Release 6.3.98 is being rolled out to the first phase of regions as listed here.
This release contains internal and customer bug fixes.
]]>Monitor semantic governance policies with built-in metrics (Preview)
Built-in Cloud Monitoring metrics for the semantic governance policy engine
are available in Preview. You can now observe request throughput, evaluation
counts, latencies, verdict distribution (ALLOW versus DENY), and LLM
token consumption for the policy engine directly in
Metrics Explorer, query them through
the Cloud Monitoring v3 API and PromQL, and use them in alerting policies.
For more information, see Monitor semantic governance policies.
Release 6.3.97 is now available for all regions.
]]>To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
To improve security, starting in August 2026, App Engine opts your application into TLS version 1.2 and later. You can opt out until the end of August 2026. Starting in September 2026, App Engine might permanently block insecure traffic with TLS version 1.1 and earlier. For more information, see Secure minimum TLS.
As detailed in our 2026 Environmental Report (p. 22), Google is now using Granular Certificates purchased from the marketplace to strategically match more of our load on an hourly basis. To accurately incorporate these certificates into the Cloud customers' allocation of carbon intensity calculations, the July 2026 semi-annual methodology refresh will be delayed by one month. We will provide further updates once the revised data is available.
The Telemetry API for logs ingestion is generally available (GA). You can ingest OTLP logs into Cloud Logging by using an OpenTelemetry Collector, an OTLP exporter, and the Telemetry API. For more information, see OTLP ingestion overview.
SapBdcProductBuilder incorrectly enforced SAP-versioned sections (ecc, s4, common) in table_settings.Gemini Enterprise: Gemini 3.7 Flash available in the mobile app
Gemini 3.7 Flash is generally available (GA) in the Gemini Enterprise mobile app. Mobile app users can select and use the Gemini 3.7 Flash model for their conversations within the app. To make the model available, administrators must turn on the Gemini 3.7 Flash feature toggle in the Google Cloud console.
For more information, see:
GKE cluster versions have been updated.
New versions available for upgrades and new clusters.
The following versions are now available for new GKE clusters, and for manual control plane upgrades and node upgrades for existing clusters. For more information about versioning and upgrades, see GKE versioning and support and About GKE cluster upgrades.
This release includes new GKE versions that use updated Container-Optimized OS images. These updated images are cumulative, incorporating security fixes from all Container-Optimized OS versions released since the previous GKE release.
To identify the specific vulnerabilities that were resolved in each updated Container-Optimized OS image, see the Security release notes for that image. The following table includes links to the release notes for each updated Container-Optimized OS image:
| GKE version | Container-Optimized OS version | Details |
|---|---|---|
| 1.31.14-gke.2579000 | cos-117-18613-675-37 | cos-117-18613-675-37 release notes |
| 1.32.13-gke.2268000 | cos-117-18613-675-37 | cos-117-18613-675-37 release notes |
| 1.33.13-gke.1462000 | cos-121-18867-528-36 | cos-121-18867-528-36 release notes |
| 1.35.7-gke.1150000 | cos-125-19216-532-62 | cos-125-19216-532-62 release notes |
| 1.37.0-gke.1173000+preview | cos-129-19506-299-60 | cos-129-19506-299-60 release notes |
[Spotlight Feature] Monitor your data latency with the Health Hub
This feature is in public preview. The Health Hub now includes two new tables to track the ingestion latency at both the source level and the log-type level. In addition, you can select a specific source or log type to open the Data Health Deep Dive page and view detailed information about ingestion latency. For more information, see Monitor health of data sources.
Key capabilities include:
You can use custom constraints with Organization Policy to provide more
granular control over specific fields for Agent Identity resources, such as
agentidentity.googleapis.com/AuthProvider. For more information, see
Use custom organization policies for Agent Identity.
This feature is in
GA.
On August 13th, 2026, we began maintenance updates of Apigee instances configured for maintenance windows.
If you set a preferred window for maintenance for your instance, and your instance version is below 1-18-0-apigee-2, your instance will be updated to 1-18-0-apigee-2 within the next seven to 21 days. A notification containing the expected date of upgrade will be sent within the next two business days.
For more information on participating in scheduled maintenance windows, see Maintenance overview and Manage Apigee instance maintenance windows.
On August 13th, 2026, we released an updated version of Apigee (1-18-0-apigee-3).
| Bug ID | Description |
|---|---|
| 532147587 | To fix forward proxy support. |
| 537657987 | Fixed a bug where watcher failed to reconcile all routes if an environment was not found in the control plane. |
| 543022076 | Google Cloud BOM upgrade (protobuf 4.x, gRPC 1.81, Guava 33.5). One user-visible change: a malformed inbound gRPC request frame is now reported to the client as grpc-status INTERNAL(13) and recorded in analytics as x-apigee.grpc.status=13, where it was previously an Apigee ServiceUnavailable fault seen as UNAVAILABLE(14) with no x-apigee.grpc.status recorded. Otherwise no user facing impact, but any prod issue related to gcp, protobuf or gRPC may relate to this. |
| 542242046 | Fixed LLMTokenQuota metering the request against an arbitrary quota bucket when the API Product declared multiple models and the request carried no model. |
| 531731614 | Apigee analytics fields ai_llm_response_token_count, ai_llm_prompt_token_count, ai_llm_model_name, and ai_llm_model_provider are available in the Custom Report when LLMTokenQuota and PromptTokenLimit policies are used in Apigee proxies. |
| 492044413 | LLMTokenQuota resolves the model from the API Product LLM Operation when LLMModelSource is omitted and the request body has no model field. |
| 67169710 | Adds an opt-in <DynamicClientIdSupported> boolean XML element to the OAuthV2 policy. When true, AbstractOAuthStepExecution.extractClientDetails() preserves any non-empty ClientID/ClientSecret already present on the OAuthClientContext. |
| 531731614 | Apigee auto identifies the providers and publishes them to analytics. |
| 537396574 | Added feature to rotate the apigee-ca certificate. |
| 540861752 | Aligned the ApigeeDeployment conversion hub with its v1alpha3 storage version. Internal change; no effect on existing ApigeeDeployment resources. |
| 540861752 | Aligned the ApigeeDeployment custom resource's conversion hub with its v1alpha3 storage version. This internal change does not affect existing ApigeeDeployment resources. |
| N/A | Updates to infrastructure and libraries. |
| Bug ID | Description |
|---|---|
| 535928300 | Security fix for Apigee. Fixed a security issue in JWT refresh token revocation handling. |
| 539515020 | Security fix for Apigee. Fixed a security issue in the MessageValidation policy. |
| 535928530 | Security fix for Apigee. Fixed a security issue in the OAuthV2 policy. |
| 535683286 | Security fix for Apigee. Fixed a security issue in HTTP target interim-response handling. |
| N/A | Security fix for Apigee infrastructure. |
BigQuery pipelines now support automated metadata enrichment, allowing you to define semantic metadata directly in SQLX configurations for synchronization with Knowledge Catalog. In addition, the Data Engineering Agent proactively generates this metadata based on your intent or context. For more information, see Metadata enrichment and data quality scorecard integration. This feature is in Preview.
Cloud SQL for MySQL 9.7 is generally available (GA).
For more information about the differences between MySQL 9.7 and previous versions, review the MySQL 9.7 Release Notes.
By default, if you specify MySQL 9.7 as the version when you create a Cloud SQL instance (either primary or replica) using the gcloud CLI, the Google Cloud console, or the REST API, then the Cloud SQL edition is Cloud SQL Enterprise Plus edition.
This release introduces support for key MySQL 9.7 capabilities, including:
Additionally, this release introduces several database flag changes.
| MySQL database flag | Action | Details |
|---|---|---|
activate_mandatory_roles |
Added / Supported | Exposes mandatory roles configuration. Default is ON. |
innodb_native_foreign_keys |
Added / Supported | Exposes SQL layer foreign key handling configuration. |
table_open_cache_triggers |
Added / Supported | Configures trigger cache size limits. |
connection_memory_status_limit |
Added / Supported | Sets limit configurations on connection memory. |
global_connection_memory_status_limit |
Added / Supported | Sets global limit configurations on connection memory. |
performance_schema_max_logger_classes |
Added / Supported | Sets limits for performance schema logger classes. |
caching_sha2_password_proxy_users |
Added / Supported | Configures proxy users support for caching SHA-2. |
caching_sha2_password_enforce_storage_format |
Added / Supported | Enforces storage format rules for caching SHA-2. |
caching_sha2_password_storage_format |
Added / Supported | Sets storage format defaults for caching SHA-2. |
innodb_log_file_size |
Removed | Flag not supported in MySQL 9.7 and later. |
innodb_log_files_in_group |
Removed | Flag not supported in MySQL 9.7 and later. |
innodb_undo_tablespaces |
Removed | Flag not supported in MySQL 9.7 and later. |
mysql_native_password_proxy_users |
Removed | Flag not supported in MySQL 9.7 and later. |
replica_parallel_type |
Removed | Flag not supported in MySQL 9.7 and later. |
slave_parallel_type |
Removed | Flag not supported in MySQL 9.7 and later. |
temptable_use_mmap |
Removed | Flag not supported in MySQL 9.7 and later. |
For more information about MySQL database flags, see Configure database flags.
The following features are rolling out over the next week.
Fullscreen charts
You can view individual charts in fullscreen mode. Click the fullscreen button in the chart header to expand the chart. This feature is not available for scorecards and gauge charts.
Rotate components
You can rotate text boxes, images, and shapes in Data Studio. Report creators can rotate these components on a non-responsive canvas and can reset the rotation to 0 degrees.
Center labels on stacked bar charts
You can position labels in the center of stacked bar charts. If insufficient space is available to center the label within the bar, the label is displayed outside the bar.
For more information, see the Bar chart and column chart reference.
Search for settings
You can search for settings in the Setup and Style tabs of the properties panel.
Copy chart as image
You can copy a chart as a PNG image to your clipboard.
Bubble chart border color
You can modify the border color of bubbles in bubble charts.
Dataform workflows and BigQuery pipelines now support automated metadata enrichment for BigQuery tables and views, allowing you to define semantic metadata directly in SQLX configurations for synchronization with the Knowledge Catalog. For more information, see Add metadata for Knowledge Catalog. This feature is in Preview.
Error Reporting can report stack traces collected from Rust applications
using std::backtrace.
To enable, set the RUST_BACKTRACE=1 environment variable and
make sure debug symbols are enabled.
For more information, see ReportedErrorEvent.
Gemini Enterprise: Create, upload, and share custom skills
Skills are reusable custom instructions that help the Gemini Enterprise assistant perform specific tasks. End users can create, upload, and share skills. To enable this feature, administrators must turn on the skills and skill-sharing settings in Feature Management. Administrators can also configure skill availability and approve skill-sharing requests.
This feature is generally available (GA). For more information, see the following:
Gemini Enterprise: Use Gemini 3.7 Flash
Gemini 3.7 Flash is generally available (GA) in the global, us, and eu
regions. To make Gemini 3.7 Flash available to users in the Gemini Enterprise
app, administrators must turn on the Gemini 3.7 Flash feature toggle in the
Google Cloud console.
For in-country regions where the model isn't supported, administrators
can still enable the model by confirming a warning that traffic is routed
to the global endpoint, which doesn't support regional data residency.
Gemini 3.7 Flash is also available in Agent Designer workflow agents. Updates take up to a day to appear in workflow agents.
For more information, see:
Gemini Enterprise: Gemini 3.7 Flash availability in mobile app
Gemini 3.7 Flash is not available in the Gemini Enterprise mobile app. A new release note will be added when the Gemini 3.7 Flash rollout for the Gemini Enterprise mobile app is complete.
Gemini 3.7 Flash is generally available
Gemini 3.7 Flash is now generally available (GA) and available for production use. This model is our first model to enable agentic video processing enabled by default.
For more information on 3.7 Flash, see the model page.
Scheduled Maintenance
SOAR database and infrastructure maintenance is scheduled to take place during the standard maintenance window on Sunday, August 16. During this window, your system will experience a brief period of downtime. No customer action is required.
Scheduled Maintenance
SOAR database and infrastructure maintenance is scheduled to take place during the standard maintenance window on Sunday, August 16. During this window, your system will experience a brief period of downtime. No customer action is required.
You can migrate workloads from self-managed Redis and Valkey instances running on Google Cloud to Memorystore for Valkey. This feature is Generally Available.
AI Protection supports data residency in the Kingdom of Saudi Arabia (KSA) for all Security Command Center service tiers.
For more information, see Planning for data residency.
The ANTHROPIC_API_KEY, GEMINI_API_KEY, and OPENAI_API_KEY infoType detectors are available in all regions. For more information about all built-in infoTypes, see the InfoType detector reference.
VPC Service Controls feature (Status: Preview): Support for optimizing service perimeters using the VPC Service Controls recommender is available.
The recommender detects architectural risks and perimeter misconfigurations, including the following:
Critical resources at risk of exfiltration: Identifies active and sensitive services (such as BigQuery and Cloud Storage) operating outside service perimeters.
Unconfigured VPC accessible services: Identifies perimeters that leave APIs unrestricted from within the security boundary.
Misconfigured VPC accessible services: Identifies mismatches between allowed accessible APIs and restricted services inside a perimeter.
For more information, see Optimize perimeters with recommender.
Agent Search: Gemini 3.5 Flash answer generation
You can generate answers with the Gemini 3.5 Flash model.
For more information, see Answer generation model versions and lifecycle and Gemini 3.5 Flash.
]]>Configure and deploy MCP servers with gcloud CLI
You can use the gcloud apihub locations configure-and-deploy-server command to configure and deploy API hub Model Context Protocol (MCP) servers to an attached Apigee runtime.
Define MCP tools inline or by referencing a YAML or JSON specification file to expose your API hub operations for agent integrations.
For more information, see gcloud CLI for API hub.
Table Explorer behavior has moved to the Reference panel. Table Explorer has been deprecated. For more information, see "Use the Reference panel" in Run a query.
You can use parameterized views in Bigtable to dynamically filter data ranges for logical views based on application context and mitigate SQL injection risks. This feature is generally available (GA). For more information, see Parameterized views overview.
You can use the CLUSTER_ATTRIBUTE() filter to restrict continuous materialized
view processing to specific clusters. This function lets you isolate views
within an instance. This feature is generally available (GA).
For more information, see Non-deterministic SQL functions.
The following remote MCP servers automatically generate a trace span for
tools/call operations. These spans can help you understand the behavior of
your agentic applications. For more information, see
Investigate MCP calls using Trace.
Google Cloud Observability automatically generates trace exemplars for charts on custom dashboards that display the result of a SQL query when the query runs against your trace data and satisfies some constraints. The exemplars link the SQL query result to specific traces. This feature is in Preview.
For more information, see Generate and display trace exemplars.
Cloud Workstations supports Compute Engine suspend and resume in Preview. You can configure workstation VMs to suspend when they reach their idle timeouts, referred to as auto-sleep in the Google Cloud Console, rather than shutting down and deleting the VM, by using the IdleAction workstation configuration setting.
Gemini Enterprise: GitHub connector with data federation
The GitHub connector with data federation is generally available (GA) in Gemini Enterprise. The connector lets you search and act on GitHub repositories, issues, and pull requests directly from the Gemini Enterprise agent, with tool actions such as creating branches, adding issue comments, merging pull requests, and pushing files.
For more information, see the Connect GitHub with data federation documentation.
Gemini Enterprise: AlphaEvolve HPC solution
The AlphaEvolve HPC solution provides a distributed, containerized infrastructure for running large-scale evolutionary code optimization experiments on Google Cloud. If your evaluations require specialized hardware or exceed the resource limits of a single machine, use the AlphaEvolve HPC solution.
For more information, see AlphaEvolve for HPC use cases.
CodeMender CLI: Sandbox enabled by default
This release updates the CodeMender CLI default behavior:
config.yaml, by passing --sandbox=false to CLI commands, or bypass it using the --unrestricted flag.For more information, see Install the CLI and configure.
Google Cloud CCaaS prerelease notes
Here are the pre-release notes for the next version of Google Cloud CCaaS. When we release this version, we expect the new capabilities to be as shown here.
Agent desktop supports parameters in custom panel URLs
In the agent desktop, you can now configure fixed and dynamic parameters to include in the URLs of custom panels. This lets you pass relevant session, agent, and customer context into custom panels.
Callback offer restrictions
You now have greater control over when callbacks are offered. You can configure the following:
Prevent callback offers from being made outside of callback hours.
Prevent callback offers that will likely occur outside of callback hours. If conditions improve (that is, EWT decreases), the system adjusts and can offer callbacks.
Administrators: We've added the following checkboxes to the CCAI Platform portal:
Restrict callback offer outside of callback window
Restrict callback offer that will exceed hours of operation. If queue condition improve offer callbacks
These checkboxes are available in the following locations:
The Settings > Call > Callback Settings pane (to configure globally).
The Settings > Queue > IVR (Interactive Voice
Response) > Edit / View >
QUEUE_NAME > Callback Settings
> Configure > Callback Management pane (to configure
a queue).
This release addresses the following issues:
Fixed an issue where chat transcripts incorrectly displayed undefined joined instead of the agent's name during a chat transfer when real-time redaction was enabled.
Fixed an issue where agents were incorrectly placed into Unresponsive status after answering a call if the customer declined the system's callback attempt.
Fixed an issue where overcapacity deflection didn't trigger when agent-initiated outbound or direct-inbound calls were transferred to a queue, causing users to wait indefinitely.
Fixed an issue where the Answer button didn't appear for incoming calls in the agent desktop, preventing agents from accepting calls.
Fixed an issue where live translation didn't activate (or translated in the wrong direction) after a chat was transferred between queues.
Fixed an issue where predictive campaign calls became stuck in a queued state, causing agents to appear available despite being unable to receive new calls.
Fixed an issue where web chats became stuck in a queued state and were never assigned to an agent.
Fixed an issue where previously closed chat sessions briefly reappeared and gained focus when launching agent desktop.
Fixed an issue where agents couldn't submit disposition codes and notes during wrap-up.
Fixed an issue where agents couldn't change their status after a call ended abnormally, causing them to remain stuck in the wrap-up state.
Fixed an issue where the Call History list incorrectly displayed the same customer phone number for all Acqueon campaign calls.
Fixed an issue where an agent's status incorrectly remained Available during outbound calls and wrap-up periods, allowing the routing engine to offer new inbound calls to occupied agents.
Fixed an issue where agents were forced to re-authenticate when opening the chat adapter or email adapter despite having an active session.
Fixed an issue where the text screen in the email adapter suddenly re-rendered while typing, causing characters to disappear or be displaced.
Fixed an issue where the headings in Generative AI session summaries in chat wrap-up notes lost their bold formatting when saved.
Fixed an issue where the global after-hours deflection message played despite a queue-level custom redirect to a phone number being configured.
Fixed an issue where calls abandoned by a customer while waiting in a transfer queue were incorrectly reported as queue failures.
Fixed an issue where transient CRM errors caused significant delays in retrieving ticket IDs during active calls.
Fixed an issue where the CCAI Platform portal incorrectly displayed a chat status as Unknown (instead of Check In Timeout) when a consumer didn't check in.
Fixed an issue where the call adapter incorrectly displayed Portuguese BR instead of Portuguese (Portugal) during call handoffs.
Fixed an issue where agents making outbound calls remained in an Available status, which incorrectly allowed new inbound calls to be routed to them while they were already engaged.
Fixed an issue where Agent Assist live transcription and generative summary stopped working mid-call following a transfer or a hold-and-resume cycle.
Fixed an issue where the transfer menu delayed loading queues.
Fixed an issue where the agent adapter incorrectly displayed an agent's status as Unavailable when a custom status, such as Break or Special Task, was selected.
Fixed an issue where chats escalated from a virtual agent to a human agent queue bypassed menu-level after-hours and over-capacity deflection messages.
Fixed an issue where the call adapter defaulted to English in the Outbound call screen regardless of the agent's system language.
Fixed an issue where Call and Chat each appeared twice in the Dashboard menu when using high browser zoom levels or small window resolutions.
Fixed an issue where legacy dashboards were restricted to English-only labels.
Fixed an issue where users in SAML-only or SSO-enabled environments received invitation emails directing them to a non-existent Forgot Password flow.
Fixed an issue where the user activity logs incorrectly recorded an end-user ID instead of the agent's ID when a chat disconnected.
Fixed an issue where customer calls were abandoned during payment transactions when DTMF inputs were provided.
Fixed an issue where agents could see and select outbound caller IDs that weren't assigned to their teams or queues.
Fixed an issue where post-session virtual agent transfers stalled if a customer left the chat while still in a queue for a human agent.
Fixed an issue where agents and end-users could hear each other's voices despite the agent putting the call on hold.
Fixed an issue where the agent adapter incorrectly reverted to displaying English when initiating an outbound call in a non-default language.
Fixed an issue where the Hide Agent Assist button wasn't appearing in the user interface during the call disposition phase.
Fixed an issue where cascade agent availability conditions weren't enforced, causing regional agents to be incorrectly routed into international queues and leaving local queues understaffed.
Fixed an issue where agents appeared available but weren't receiving calls.
Fixed an issue where administrators received multiple email notifications instead of a single notification after deactivating a call channel.
Fixed an issue where Telnyx error handling was incorrectly configured, causing system alerts to fail during call disconnect or hold actions.
Fixed an issue where inbound IVR calls were stuck in a queued state if a caller disconnected during initial call processing.
Fixed an issue where callbacks became permanently stuck in a queued state if an agent missed a projected call.
Fixed an issue where calls using Telnyx or Nexmo numbers received an "An Application error has occurred" message during overcapacity deflection or automatic redirection.
Fixed an issue where call recordings weren't delivered or processed correctly.
Fixed an issue under Settings > Operation Management > Localization > Manage Location Setting where configured locations with Portuguese (Portugal) or Spanish (Spain) selected incorrectly displayed as Unknown in the Language column, and locations with Spanish (Mexico) selected incorrectly displayed as Spanish (Spain).
Fixed an issue in iOS and Android SDKs where the end user's initial message and custom data weren't correctly passed to the routing API during the chat menu fetch.
Fixed an issue where inbound calls cancelled by the caller within the first second incorrectly remained active in the system for several hours.
Fixed an issue where custom SIP headers were dropped. This occurred when a directly dialed agent was over capacity, the call was redirected to the agent's queue, and that queue was configured to redirect to a SIP URI.
Fixed an agent desktop issue where the chat screen went blank when an agent accepted or dismissed a chat.
Fixed an issue where the Deflections - Calls dashboard incorrectly reported the queues that calls were redirected to when using percent allocation.
Fixed an issue where auto-answered interactions became stuck in a Queued status despite being accepted by an agent.
Fixed an issue where the chat adapter in a CRM integration didn't post outbound messages when using rich text formatting.
Fixed an issue where the agent_activity_logs endpoint experienced timeouts
and degraded performance when processing large data requests.
Fixed an agent desktop issue where outbound calls appeared to start successfully but didn't connect to the end-user.
Fixed an issue where New photo received notifications appeared whenever an agent switched between chat tabs.
Fixed an issue where short outbound calls incorrectly displayed a This call was abandoned by the customer message.
Fixed an issue where agents experienced significant delays when switching between multiple concurrent web chats.
Fixed an issue where attaching PDF or text files in the chat adapter failed or timed out.
Fixed an issue where calls transferred from a virtual agent to an agent extension were incorrectly deflected.
Fixed an issue where reordering queues in the CCAI Platform portal experienced extreme latency and didn't update visually without a manual page refresh.
Fixed an issue where chat history for added agents became unavailable after a page refresh.
Fixed an issue where calls to the user_activity_logs endpoint experienced
significant delays.
Fixed an issue where repeated voicemail re-reads caused duplicate participant entries and inflated session data.
Fixed an issue where callback sessions remained in a Connected status indefinitely after completion.
Fixed an issue where agents using IdP-initiated SAML SSO were automatically routed to the default home page instead of the agent desktop.
Fixed an issue where DTMF options in the softphone didn't register with external IVR systems during outbound calls.
Fixed an issue where agents were unable to receive or fetch incoming calls.
Fixed an issue where supervisors using Telnyx who ended a monitoring session were unable to monitor subsequent calls.
Fixed an issue where empty chat bubbles appeared in the chat adapter when an end-user used suggestion chips to respond to a virtual agent.
Fixed an issue where duplicated contact handle-duration events caused inaccurate reporting for call and chat interactions.
Fixed an issue where agents didn't receive the correct error messages when their microphone was disabled or inaccessible during a call.
Fixed an issue where incoming calls were incorrectly multicasted and didn't auto-answer when an agent was already in an active chat session.
Fixed an issue where calls deflected to external SIP or phone destinations
were missing the ends_at timestamp in call data.
Fixed an issue where the chat timeout event wasn't correctly emitted to the headless web SDK, causing sessions to remain active for up to 120 minutes, regardless of the configured settings.
Fixed an issue where agents using Salesforce-Lightning or Zendesk embeds couldn't maintain a stable presence connection.
Fixed an issue where headless web SDK client methods didn't work after a mid-session authentication update.
Fixed an issue where agents were unexpectedly logged out of all active sessions.
Fixed an issue where calls producing multiple recording segments resulted in duplicate entries in Customer Experience Insights.
Fixed an issue where the parent_id field was missing from callback call
responses in the manager API.
Fixed an issue where virtual agent transcripts were lost during escalations to live agents.
Fixed an issue in the agent desktop where the sentiment score appeared in the Call details panel despite sentiment analysis being turned off in the conversation profile.
Fixed an issue where the Escalated To Language column in the
Escalations table of the Virtual Agent - Calls dashboard incorrectly
displayed Unknown for French (Canada) calls.
Fixed an issue where calls didn't advance to the next cascade group after the timer threshold was reached.
GKE cluster versions have been updated.
New versions available for upgrades and new clusters.
The following versions are now available for new GKE clusters, and for manual control plane upgrades and node upgrades for existing clusters. For more information about versioning and upgrades, see GKE versioning and support and About GKE cluster upgrades.
This release includes new GKE versions that use updated Container-Optimized OS images. These updated images are cumulative, incorporating security fixes from all Container-Optimized OS versions released since the previous GKE release.
To identify the specific vulnerabilities that were resolved in each updated Container-Optimized OS image, see the Security release notes for that image. The following table includes links to the release notes for each updated Container-Optimized OS image:
| GKE version | Container-Optimized OS version | Details |
|---|---|---|
| 1.31.14-gke.2543000 | cos-117-18613-675-28 | cos-117-18613-675-28 release notes |
| 1.32.13-gke.2231000 | cos-117-18613-675-28 | cos-117-18613-675-28 release notes |
| 1.33.13-gke.1414000 | cos-121-18867-528-21 | cos-121-18867-528-21 release notes |
| 1.34.10-gke.1079000 | cos-125-19216-532-42 | cos-125-19216-532-42 release notes |
| 1.35.7-gke.1027000 | cos-125-19216-532-25 | cos-125-19216-532-25 release notes |
| 1.36.3-gke.1244000 | cos-129-19506-299-60 | cos-129-19506-299-60 release notes |
[Spotlight Feature] Analyze feed activity with Cloud Logging
This feature is in public preview. To use this feature, your Google SecOps instance must be configured with a Bring Your Own Project (BYOP) Google Cloud project. You can now monitor, debug, and troubleshoot Google SecOps ingestion pipelines and feeds using Cloud Logging. By sending, viewing, and querying ingestion and feed activity logs in Logs Explorer, you can diagnose log delivery issues, such as, missing, delayed, or failing logs, and decrease the time required to resolve ingestion anomalies.
This visibility into push- and pull-based ingestion mechanisms provides the following capabilities:
For more information, see Analyze feed activity with Cloud Logging.
CyberArk Credential Provider: Version 5.0
The following new job has been added:
Microsoft Graph Mail: Version 45.0
The following new actions have been added:
Microsoft Graph Mail Delegated: Version 22.0
The following new actions have been added:
Active Directory: Version 45.0
Fixed an issue in the following action where entity properties were incorrectly reset on update:
AWS WAF: Version 14.0
Cisco Umbrella: Version 21.0
Fixed an issue in the following action where entity attachment failed due to a bytes object serialization error:
CrowdStrike Falcon: Version 81.0
Added the ability to use device IDs as input parameters in the following actions:
Enrichment
Fixed an issue in the following action where unsupported entity types were selected during enrichment:
GitSync
Fixed an issue in the following action where the Include Playbook Blocks parameter was ignored when a folder allowlist was used:
Microsoft 365 Defender: Version 30.0
Added support for GCC High tenants by dynamically constructing API token scopes and adding a configurable API Root parameter in the following connector:
Improved error handling and alert processing mechanisms in the following job:
Microsoft Graph Mail: Version 45.0
Fixed an issue in the following action where an unhandled exception occurred when a user mailbox was not found:
Microsoft Graph Mail Delegated: Version 22.0
Fixed an issue in the following action where an unhandled exception occurred when a user mailbox was not found:
[Spotlight Feature] Analyze feed activity with Cloud Logging
This feature is in public preview. To use this feature, your Google SecOps instance must be configured with a Bring Your Own Project (BYOP) Google Cloud project. You can now monitor, debug, and troubleshoot Google SecOps SIEM ingestion pipelines and feeds using Cloud Logging. By sending, viewing, and querying ingestion and feed activity logs in Logs Explorer, you can diagnose log delivery issues, such as, missing, delayed, or failing logs, and decrease the time required to resolve ingestion anomalies.
This visibility into push- and pull-based ingestion mechanisms provides the following capabilities:
For more information, see Analyze feed activity with Cloud Logging.
The workflow for creating workforce identity pool providers in the Google Cloud console changed. After submitting the initial provider configuration, the console directs you to a centralized page to configure provider attributes, including attribute mappings, attribute conditions, and extra attributes.
For more information, see Manage workforce identity pools and providers.
Key updates in these runtime versions include:
2.3 runtime:
1.49 to support lineage for tables created using the Lakehouse Runtime catalog.Connectivity Tests supports using a Cloud Run job as a source endpoint for connectivity testing.
For more information, see Test from a Cloud Run job to a destination.
]]>On August 11, 2026 we released an updated version of the Apigee hybrid software, v1.16.9.
| Bug ID | Description |
|---|---|
| 514973778 | Fixed an issue where the SanitizeUserPrompt and SanitizeModelResponse policies failed to tolerate unknown fields while parsing responses from the Model Armor Service. |
| 543171828 | Fixed an issue where the apigee-logger DaemonSet failed to schedule on cluster nodes without custom node labels due to a default logger.nodeSelector in the Helm chart. |
Various security and CVE fixes are included in this release.
Query templates for data clean rooms are generally available (GA). Query templates allow data clean room owners and publishers to share predefined queries without exposing the underlying tables and views.
Additionally,
table parameters
in table-valued functions (TVFs) are
generally available
(GA). You can use the ANY TABLE type as a table parameter to create
generic functions that accept tables of any structure.
Cloud Run NVIDIA L4 GPU driver version 580.x.x is available for services, jobs, and worker pools.
Generally available: Compute flexible committed use discounts (CUDs) are available for G2 and G4 GPU accelerator-optimized machine series. The supported resources include vCPUs, memory, Local SSD disks, and GPUs.
Compute flexible CUDs are spend-based CUDs that apply to eligible Google Cloud spend across Compute Engine, GKE, and Cloud Run. For G2 and G4 machine series, compute flexible commitments provide the flexibility to switch between eligible machine series and regions depending on your workload needs. For GPUs that belong to these machine series, compute flexible commitments don't require attached reservations.
For more information, see Compute flexible CUDs.
A vulnerability (CVE-2026-6726) in the Trusted Computing Group's TPM 2.0 reference implementation code was discovered and is being addressed. For more information, see the GCP-2026-054 security bulletin.
A vulnerability affecting Intel TDX firmware was discovered and is being addressed. For more information, see the GCP-2026-053 security bulletin.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.18.39 | v29.4.3 | v2.3.2 | See List |
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.12.94 | v27.5.1 | v2.2.6 | See List |
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.18.41 | v29.4.3 | v2.3.2 | See List |
Added support for installing the Vast 4.5.8 NFS client drivers with cos-dkms.
Added kernel patch to reduce bcache garbage collection sleep interval to prevent I/O stalls.
Added support for installing the Vast 4.5.8 NFS client drivers with cos-dkms.
Added kernel patch to reduce bcache garbage collection sleep interval to prevent I/O stalls.
Fixed CVE-2026-33186 in google-guest-agent.
Added kernel patch to reduce bcache garbage collection sleep interval to prevent I/O stalls.
Mask nfttables-restore.service to address time to ssh regression.
Fixed CVE-2026-64227 in the Linux kernel.
Mask nfttables-restore.service to address time to ssh regression.
Fixed KCTF-8173f7e in the Linux Kernel.
Fixed CVE-2026-64279 in the Linux kernel.
Updated app-admin/node-problem-detector to v0.8.25.
Fixed CVE-2026-64286 in the Linux kernel.
Fixed KCTF-8173f7e in the Linux Kernel.
Fixed CVE-2026-64287 in the Linux kernel.
Upgraded net-libs/nghttp2 to 1.69.0 and fixed CVE-2026-58055.
Fixed CVE-2026-64352 in the Linux kernel.
Fixed CVE-2026-64375 in the Linux kernel.
Fixed CVE-2026-64401 in the Linux kernel.
Fixed CVE-2026-64413 in the Linux kernel.
Fixed CVE-2026-64416 in the Linux kernel.
Fixed CVE-2026-64476 in the Linux kernel.
Fixed CVE-2026-64508 in the Linux kernel.
Fixed CVE-2026-64530 in the Linux kernel.
Fixed CVE-2026-64532 in the Linux kernel.
Fixed CVE-2026-64533 in the Linux kernel.
Fixed CVE-2026-64534 in the Linux kernel.
Fixed CVE-2026-64535 in the Linux kernel.
Fixed CVE-2026-64538 in the Linux kernel.
Fixed CVE-2026-64542 in the Linux kernel.
Fixed CVE-2026-64545 in the Linux kernel.
Fixed CVE-2026-64546 in the Linux kernel.
Fixed CVE-2026-64548 in the Linux kernel.
Fixed CVE-2026-64552 in the Linux kernel.
Fixed CVE-2026-64554 in the Linux kernel.
Fixed CVE-2026-64555 in the Linux kernel.
Fixed KCTF-8173f7e in the Linux Kernel.
Runtime sysctl changes:
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.6.143 | v27.5.1 | v2.0.10 | See List |
Update dev-lang/go to 1.25.12.
Fixed CVE-2026-64279 in the Linux kernel.
Fixed CVE-2026-64319 in the Linux kernel.
Fixed CVE-2026-64352 in the Linux kernel.
Fixed CVE-2026-64375 in the Linux kernel.
Fixed CVE-2026-64401 in the Linux kernel.
Fixed CVE-2026-64413 in the Linux kernel.
Fixed CVE-2026-64474 in the Linux kernel.
Fixed CVE-2026-64476 in the Linux kernel.
Fixed CVE-2026-64535 in the Linux kernel.
Fixed KCTF-8173f7e in the Linux Kernel.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.6.143 | v24.0.9 | v1.7.34 | See List |
Update dev-lang/go to 1.25.12.
Fixed CVE-2026-64279 in the Linux kernel.
Fixed CVE-2026-64319 in the Linux kernel.
Fixed CVE-2026-64352 in the Linux kernel.
Fixed CVE-2026-64375 in the Linux kernel.
Fixed CVE-2026-64401 in the Linux kernel.
Fixed CVE-2026-64413 in the Linux kernel.
Fixed CVE-2026-64535 in the Linux kernel.
Fixed CVE-2026-64548 in the Linux kernel.
Fixed CVE-2026-64556 in the Linux kernel.
Fixed KCTF-8173f7e in the Linux Kernel.
google-auth, google-cloud-bigquery, google-cloud-dataform, google-cloud-resource-manager, google-cloud-service-usage and google-cloud-storage.Firestore now supports the asia-southeast3 Bangkok region.
For a full list of supported locations, see Locations.
Firestore in Datastore mode (Datastore) now supports the asia-southeast3 Bangkok
region.
For a full list of supported locations, see Locations.
Gemini Enterprise: New data stores and support for new actions (Public Preview)
The following data stores are available in Public Preview in Gemini Enterprise:
You can search and read data from these data stores using natural language.
Additionally, the following data stores support new actions in Public Preview:
Gemini Enterprise: Manage overages, spend limits, and costs for invoiced Cloud Billing accounts
If your project has an invoiced Cloud Billing account and at least one active, non-free-trial subscription, administrators can enable overages, configure monthly spend limits, and monitor feature usage and costs in Gemini Enterprise:
For more information, see:
Google Distributed Cloud (software only) for VMware 1.35.400-gke.81 is now available for download. To upgrade, see Upgrade clusters. Google Distributed Cloud 1.35.400-gke.81 runs on Kubernetes v1.35.3-gke.400.
If you use a third-party storage vendor, check the listing of our previously-qualified storage partners.
After a release, it takes approximately 7 to 14 days for the version to become available for use with GKE On-Prem API clients: the Google Cloud console, the gcloud CLI, and Terraform.
The following issues were fixed in 1.35.400-gke.81:
Reconciling state
after an admin cluster upgrade. The admin cluster controller skipped
reconciling legacy cluster lifecycle components during upgrades unless an initial
migration annotation was set. If legacy user clusters still existed on the
admin cluster, missing legacy API discovery (cluster.k8s.io/v1alpha1) caused
controller reconciliation to stall. With this fix, the controller preserves
legacy components as long as any legacy user clusters exist, and prunes them only
after all user clusters have migrated to advanced clusters.
gkectl prepare failed with a permission denied error
when authenticating against a private container registry.
Google Distributed Cloud (software only) for bare metal 1.35.400-gke.81 is now available for download. To upgrade, see Upgrade clusters. Google Distributed Cloud for bare metal 1.35.400-gke.81 runs on Kubernetes v1.35.3-gke.400.
After a release, it takes approximately 7 to 14 days for the version to become available for installations or upgrades with the GKE On-Prem API clients: the Google Cloud console, the gcloud CLI, and Terraform.
If you use a third-party storage vendor, check the listing of our previously-qualified storage partners.
The following issues were fixed in 1.35.400-gke.81:
BigQuery now supports the gemini-3.1-flash-lite and gemini-3.5-flash GA
models, which are available for the us, eu, and global multi-regional
endpoints. You can use these models in all generative AI functions.
For information about how to specify a multi-regional endpoint and
how endpoints are selected, read about
locations
in the generative AI overview.
Starting September 15, 2026, the App Topology API transitions to a usage-based billing model that includes a daily free data usage allotment. For more information, see App Topology pricing.
The Cloud Run functions upgrade tool is in General Availability. Use this tool to upgrade 1st gen functions to Cloud Run functions.
A new Confidential Space image (260701) is available.
When configuring Eventarc triggers for Cloud Run destinations (including Cloud Run functions), you can specify a single delivery attempt with no retries. For more information, see Retries for Cloud Run destinations.
Various bug fixes and minor product enhancements.
Gemini Enterprise: Export user data to a CSV file
Gemini Enterprise administrators can export user data to a comma-separated values (CSV) file to sort, filter, and analyze records offline. This feature is generally available (GA).
For more information, see Export user data.
Google Cloud CCaaS 6.0
We've released version 6.0 of Google Cloud CCaaS.
Version 6.0 updates internal Google Cloud CCaaS infrastructure. It contains no customer-facing changes from version 5.2.
The timing of the update to your instance depends on the deployment schedule that you have chosen. For more information, see Deployment schedules.
A new Managed Service for Apache Airflow release has started on August 10, 2026. Get ready for upcoming changes and features as we roll out the new release to all regions. This release is in progress at the moment. Listed changes and features might not be available in some regions yet.
(Managed Airflow Gen 3) Adjusted the formula used to calculate the number of Airflow web server workers based on allocated CPU and memory resources. This update aligns with resource consumption changes in recent Airflow versions, improving web server stability.
(Airflow 3.2.2 and 2.11.1) The [api]rbac_bindings Airflow configuration
option is blocked and it isn't possible to override its value.
New Airflow builds are available in Managed Airflow (Gen 3):
New images are available in Managed Airflow (Gen 2):
Managed Service for Apache Spark (formerly Dataproc on Compute Engine):
A critical bug related to Conda channels has been fixed in-place in image versions
1.4.81, 1.5.92, 2.1.117, and 2.2.85. These image versions were released without pre-configured Conda channels.
Required customer actions: To comply with Google requirements, recreate the following resources if they were created using these image versions on or before August 10, 2026:
You can use the Google Cloud console to secure access to your instances by using basic token-based authentication. This feature is available in Preview.
The integration of Security Command Center with Application Design Center for application lifecycle security assessments is generally available (GA). Design-time findings are sent to Security Command Center on demand during deployment. This feature lets you filter findings by App Hub application at the app-enabled folder and project levels.
For more information, see Application lifecycle security assessments.
Vulnerability Assessment for Google Cloud is available in General Availability.
Dynamic channel pooling (DCP) for gRPC channels in the Spanner Go and Java client libraries is generally available (GA).
DCP prevents performance issues from under-provisioned or over-provisioned channels and reduces configuration overhead. DCP is disabled by default.
For more information, see Configure the number of sessions and gRPC channels in the pools.
]]>Installed latest packages from upstream dependencies.
Installed latest packages from upstream dependencies.
Fixed the Git panel's grayed out buttons which were disabled due to an issue with the Jupyter Lab's Git plugin introduced in version 0.54.0.
Installed latest packages from upstream dependencies.
Fixed the Git panel's grayed out buttons which were disabled due to an issue with the Jupyter Lab's Git plugin introduced in version 0.54.0.
Installed latest packages from upstream dependencies.
Updated the NVIDIA GPU driver on Workbench Debian 12 images from 580.65.06 to 580.126.20 for compatibility with the Debian 12 6.1.0-52 kernel.
Installed latest packages from upstream dependencies.
Fixed the Git panel's grayed out buttons which were disabled due to an issue with the Jupyter Lab's Git plugin introduced in version 0.54.0.
Updated rich-text editor
Upgraded the rich-text editor across Google SecOps, including the Cases Wall, Use Case Upload dialog, Report Template dialog, and Dashboard Editor widget.
Key changes include:
Release 6.3.97 is being rolled out to the first phase of regions as listed here.
This release contains internal and customer bug fixes.
Updated rich-text editor
Upgraded the rich-text editor across Google SecOps, including the Cases Wall, Use Case Upload dialog, Report Template dialog, and Dashboard Editor widget.
Key changes include:
Release 6.3.96 is now available for all regions.
]]>You can now sync tables from BigQuery into your AlloyDB instance, either as a one-time operation or on a periodic schedule. This feature (in Preview) lets you enable operational analytics that benefit from low-latency, transactional access to your data lake.
For more information, see Sync BigQuery data to AlloyDB.
AlloyDB integration with BigQuery lets you connect your operational and analytical data through real-time data access (lakehouse federation), periodic data synchronization, and one-time table syncs. These features are in Preview.
For more information, see Choose how to access BigQuery data from AlloyDB.
New filter and group-by option available in Cloud Billing Reports
In Billing Reports, Cloud Billing has added the Originating products filter and Group by to provide additional options that let you analyze and understand your costs. Originating products are Google Cloud products that cause usage in another product. For example, Gemini Enterprise is an originating product when it causes usage in the Gemini Enterprise app.
To help you track and analyze your AI spend, the Originating products dimension is used in the following ways:
You can use the Originating products filter and group by option to configure your Cloud Billing report to track and analyze your Gemini Enterprise subscription and consumption costs.
The Originating products dimension supports a new preset report for quick report configuration, called Gemini Enterprise costs by SKU.
When you are viewing your costs in the Gemini Enterprise console, on the Gemini Enterprise > Usage & Spending page, the Originating products dimension supports the functionality of the costs displayed on the Gemini Enterprise Billing tab.
For more information, see the following resources:
Newly created instances configured with high availability (HA) now have Knowledge Catalog (formerly Dataplex Universal Catalog) enabled by default.
Cloud SQL for PostgreSQL instances running on PostgreSQL version 14.0 or later send updates and metadata to Knowledge Catalog in near real-time.
You can either verify enablement or disable the feature using the Google Cloud console.
For more information, see Near real-time.
Newly created instances configured with high availability (HA) now have Knowledge Catalog (formerly Dataplex Universal Catalog) enabled by default.
Cloud SQL for SQL Server instances send updates and metadata to Knowledge Catalog to help support data discovery.
You can either verify enablement or disable the feature using the Google Cloud console.
For more information, see Manage your Cloud SQL resources using Knowledge Catalog.
The accelerator-optimized g4-standard-48 machine type for securely running AI and ML workloads is generally available (GA), with the following specifications:
Resolved an issue where running the uv run cortex-build command in Windows PowerShell or Windows Command Prompt resulted in a Could not auto-import local builder warning and an Invalid builder type NoneType for category ... error.
Improved Dataform quota management in uv run cortex-deploy script.
Gemini Enterprise: Experimental agent telemetry aligned with OpenTelemetry generative AI semantic conventions
Gemini Enterprise agents emit richer, standards-aligned telemetry based on the
OpenTelemetry generative AI semantic conventions, in addition to the existing
stable telemetry. Trace spans and Cloud Logging entries include
standardized gen_ai.* attributes (for example, gen_ai.agent.name,
gen_ai.conversation.id, gen_ai.usage.input_tokens, and
gen_ai.input.messages) that describe agent, model, and tool activity.
Prompt and response message content appears in these attributes only when your observability settings allow logging of prompt inputs and response outputs; otherwise it is redacted or omitted. OpenTelemetry classifies these conventions as Development status, so this telemetry is experimental and subject to change.
For more information, see Access traces and spans.
Gemini Enterprise: Custom MCP server data stores
You can connect your custom Model Context Protocol (MCP) server with Gemini Enterprise to securely access your company's private data, custom internal tools, and MCP-compliant third-party systems.
This feature is turned off by default. To enable it, an Organization Policy Administrator must remove the organization constraint. This feature is generally available (GA).
For more information, see:
New Managed Service for Apache Spark (formerly Google Cloud Serverless for Apache Spark) subminor runtime versions:
Notes:
Apache Spark upgraded to 3.5.3 in 2.2 runtime.
Apache Gluten upgraded to 1.6 in 2.3 runtime.
Managed Service for Apache Spark latest image and runtime versions:
spark.scheduler.listenerbus.exitTimeout to 30s.Preview: Cloud KMS supports quantum-safe key import. You can use the following quantum-safe import methods:
HPKE_KEM_XWING_HKDF_SHA256_AES_256_GCMHPKE_KEM_ML_KEM_768_HKDF_SHA256_AES_256_GCMHPKE_KEM_ML_KEM_1024_HKDF_SHA256_AES_256_GCMFor more information about quantum-safe key import, see Quantum-safe key import.
Gemini Enterprise: Description field removed from Custom MCP server data store
Gemini Enterprise automatically calls your server's tools/list
endpoint to discover and understand tools, parameter schemas, and
capabilities directly. Because Gemini Enterprise routes requests using
explicit tool definitions and parameter schemas, you don't need to
write descriptions. This change simplifies data store
setup and improves routing accuracy. Existing custom MCP data stores do not
require any updates.
For more information, see Set up a custom MCP server.
Gemini Enterprise: Postponed removal of Gemini 3.5 Flash in the Global region
We have postponed the removal of Gemini 3.5 Flash from the global region
in the Gemini Enterprise app. A revised removal schedule will be provided
when available.
Advanced reporting dashboards 5.1
We've released version 5.1 of the advanced reporting dashboards.
Sub teams are included when top-level teams are selected in the Team filter
When you select a team in a Team filter on a dashboard, the data for that team's sub-teams is now included.
The Agent Metrics (Historical) Explore contains the Avg Chat Concurrency metric
The Agent Metrics (Historical) Explore now contains the Ave Chat Concurrency metric to capture concurrency trends for agents, teams, and queues.
Metrics for warm transfers and agent consultations
The Transfers - Calls and Transfers - Chats dashboards now have the following metrics in the Call Transfers and Chat Transfers tables:
Agent Connection Time (H:M:S). Reports the time spent consulting with another agent and then warm transferring to that agent.
Agent Consult Time (H:M:S). Reports the time spent consulting with another agent before returning to the call.
This release addresses the following issues:
Fixed an issue where the This Week and This Month per-hour figures on the Performance Overview and Agent Performance dashboards were inaccurate.
Fixed an issue where the Queue Abandons and Abandon % fields on the
Queue Group Performance - All and Queue Group Performance - Chats
dashboards always displayed 0 for chats.
Fixed an issue where the virtual agent call metrics on the Virtual Agent - Calls, All Interactions - Calls, and Performance Overview dashboards were incorrect.
Fixed an issue where agents appeared under Assigned Agent for calls that they weren't connected to.
Fixed an issue where exporting data from dashboards using date ranges greater than seven days failed.
Fixed an issue in the Real-time Queue Monitoring and Queued Calls dashboards where transferred calls were incorrectly displayed as Queued when they were active with an agent.
Fixed an issue where queue groups that were turned off remained visible in advanced reporting dashboards.
Fixed an issue where the Queue Time filter on the All Interactions dashboard wasn't displaying data.
Fixed an issue where the SLA Target field in the Queue Group Performance - All dashboard was empty.
Fixed an issue where the No checkbox of the Child Queues filter disappeared when cleared. No child queues is the default for this filter, so we removed the No checkbox, leaving only the Yes checkbox.
Fixed an issue where the Available Agents field overstated real capacity.
Fixed an issue on the Queue Groups - Calls and Queue Groups - Chats dashboards where the Ave Current Queue Time metric displayed incorrect data.
For time-related exports, there's now a column that displays time in HH:MM:SS format (in addition to seconds). You can't apply conditional formatting to this format in an Explore.
Fixed an issue where Max Queue Wait Time, Max Speed To Answer, and Max Queue Abandon Time displayed incorrect data in Explores.
Fixed an issue where changes to an agent's availability preferences in the agent adapter weren't reflected in the Agent Activity dashboard.
Updated the labels in the Agent Metrics (Historical) Explore to be channel-agnostic for metrics that include both calls and chats.
Looker 26.14 will roll out to Looker (original) instances on the following schedule:
Looker 26.14 is expected to include the following changes, features, and fixes.
An issue has been fixed where dashboard filters of type: string_filter could spin indefinitely if no suggestions were configured. This feature now performs as expected.
An issue has been fixed where clicking edit for a visualization from a dashboard caused the dashboard tile header to render on top of the Explore, preventing editing. This feature now performs as expected.
An issue has been fixed where LookML drill links and Liquid variables inside table cells could fail to render properly on dashboards and embedded visualizations. This feature now performs as expected.
Now generally available, the Granular Dashboard Sizing feature allows dashboard editors to change the size and layout of dashboard tiles with more granularity. Note: This item was added on August 13, 2026.
The Axis Order options in the Y tab of the visualization editor now support keyboard inputs.
Conversational Analytics data agent editors can now specify whether the agent will show its thinking or debugging information when generating a response.
A Conversational Analytics data agent now displays query results in its thinking rather than in its final response.
The timeout length for Conversational Analytics queries has been increased from two to five minutes. Note: This item was updated on August 7, 2026.
An issue has been fixed where large queries on Snowflake connections could fail with an SSLHandshakeException: No trusted certificate found error. This feature now performs as expected.
An issue has been fixed where the Show Region Field in Tooltip option was hidden for region maps that were configured without LookML map layer metadata. This feature now performs as expected.
An issue has been fixed where the list of Explores in the Conversational Analytics agent's card could be cut off for embed users with longer names. This feature now performs as expected.
An issue has been fixed where users with mixed roles across projects were unable to see and re-authorize OAuth connections for projects where they had viewer-only access. This feature now performs as expected.
An issue has been fixed where a secondary Y-axis could reset its position if a parameter value was updated. This feature now performs as expected.
An issue has been fixed where validating LookML with an empty value_format parameter could result in a 500 internal server error. This feature now performs as expected.
An issue has been fixed where modifying table calculations could cause Looker to run a new query rather than pulling results from cache. This feature now performs as expected.
An issue has been fixed where clicking the Fullscreen button while using the Conversational Analytics dashboard caused the chat view to disappear. This feature now performs as expected.
An issue has been fixed where scheduled jobs that were owned by users who authenticated by using Workforce Identity (BYOID) could fail with IAM login failures. This feature now performs as expected.
An issue has been fixed where the merged results page could return a 403 error after a period of user inactivity while cookieless embedding was being used. This feature now performs as expected.
An issue has been fixed where some characters were incorrectly displayed as HTML entities. For example," would render as ". This feature now performs as expected.
An issue has been fixed where Looker could generate incorrect SQL when creating a Self-service Explore from a spreadsheet. This feature now performs as expected.
An issue has been fixed where the drill modal incorrectly displayed a pivot option for queries that couldn't be pivoted. This feature now performs as expected.
An issue has been fixed where updating the data for a Self-service Explore could fail with a generic error. This feature now performs as expected.
An issue has been fixed where Timeline visualizations could fail to respect admin color collection overrides. This feature now performs as expected.
An issue has been fixed where visualizations were not respecting internal dashboard theme color collections, which caused custom theme color collections to revert to the defaults.
An issue has been fixed where timeout settings in the Extension SDK could be reset to their default value of 120 seconds. This feature now performs as expected.
An issue has been fixed where the items in the Axis Order section in the Y panel of the visualization edit panel were inaccessible to screen readers. This feature now performs as expected.
An issue has been fixed where non-admin user OAuth tokens remained valid even after database connection parameters were updated. This feature now performs as expected.
An issue with column numbers has been fixed that could occur when XLSX results for tiles that used pivots were generated. This feature now performs as expected.
An issue has been fixed in the Conversational Analytics interface where chats that used deleted agents or Explores could neither be deleted nor restored. This feature now performs as expected.
An issue has been fixed where searching by label in the field picker would fail to find fields where the group_label or group_item_label was set to an empty string. This feature now performs as expected.
Looker now displays a more informative error when a Conversational Analytics conversation fails to load as the result of a network failure. This feature now performs as expected.
An issue has been fixed where LookML dashboards could fail to deploy if the project name exceeded 64 characters. This feature now performs as expected.
An issue has been fixed where modifying a self-service model that was created from a blank canvas could return a Resource already exists error. This feature now performs as expected.
An issue has been fixed where Waterfall chart visualization configurations could fail to render if the X-axis label rotation was set to an empty, a null, or another undefined value. This feature now performs as expected.
An issue has been fixed where users with Admin via IAM privileges on Looker (Google Cloud core) couldn't enable the Looker Marketplace. This feature now performs as expected.
Now generally available, the Dashboard Tile Limits feature allows admins to set limits on tiles per dashboard and per tab to optimize performance, which is configured on the Content Guardrails page in the Performance Center section of the Admin panel. Note: This item was added on August 13, 2026.
The thick clone (thin clone split) feature is generally available (GA) for the Flex Unified Default-mode service level. For more information, see Manage volume clones.
The Policy Troubleshooter MCP server is generally available. To learn about using the Policy Troubleshooter MCP server to let agents and AI applications troubleshoot IAM issues and errors, see Use the Policy Troubleshooter remote MCP server.
VPC Service Controls feature: The VPC Service Controls service patterns
feature is
generally available.
You can use service patterns to explicitly configure which Google APIs (both
supported and unsupported) can be accessed from VPC networks
within a service perimeter when using the private VIP (private.googleapis.com)
or a Private Service Connect endpoint with the all-apis bundle.
For more information, see VPC Service Controls service patterns.
]]>You can now manage your Database Migration Service migration jobs directly from the destination Cloud SQL instance or AlloyDB for PostgreSQL cluster pages in the Google Cloud console. For more information, see:
The Telemetry API for metric ingestion is generally available (GA). You can ingest OTLP metrics into Cloud Monitoring by using an OpenTelemetry Collector, an OTLP exporter, and the Telemetry API. For more information, see OTLP metric ingestion overview.
Cloud Run supports sandboxes for all resources, including jobs and worker pools (Preview).
Performance capture for Cloud SQL for MySQL is now generally available (GA). Performance capture lets you take a point-in-time snapshot of your database and operating system metrics automatically and route them to Cloud Logging for root-cause analysis.
With the GA release, you can configure custom thresholds that end long-running transactions automatically before they slow down your database. In addition, the GA release includes six additional performance capture triggers:
For more information, see Cloud SQL performance capture overview.
Gemini Enterprise: Clinical Trials data store no longer available
The Clinical Trials data store is no longer available in Public Preview. The data store has returned to Private Preview.
You can generate optimized GKE configurations that can improve performance for specific workloads, such as Redis and MySQL, by using the gcloud CLI. The configurations are ConfigMaps and ComputeClasses that apply performance recommendations to the workloads and the nodes. These optimizations are available in Preview for GKE version 1.31.1-gke.12000 or later. You can measure the performance improvements by using open source benchmarks. For more information, see Optimize for workloads on GKE.
In GKE version 1.36.0-gke.3302001 and later, you can run Arm workloads on the
Autopilot container-optimized compute platform by using the general-purpose
autopilot-arm and autopilot-arm-spot ComputeClasses. You can select these
ComputeClasses in Autopilot or Standard clusters. GKE runs the workloads that
select these ComputeClasses in Autopilot mode. This compute platform improves
Pod scheduling latency, especially during autoscaling operations. For more
information, see the following documents:
Microsoft Graph Mail Delegated: Version 21.0
Updated logic for parsing email headers and S/MIME digitally signed emails in the following connector:
ServiceNow: Version 70.0
Fixed verification logic when updating reference fields in the following action:
Microsoft Graph Mail: Version 44.0
Updated logic for parsing email headers and S/MIME digitally signed emails in the following connector:
Microsoft 365 Defender: Version 29.0
Updated Google SecOps event structure, added support for incident tags and assignee filtering, updated ontology mapping, and optimized evidence handling with limits on maximum evidence items per alert in the following connector:
Jira: Version 61.0
Added support for customizable status-to-closure mapping, Case-level job scope, custom field variables in Closed Reason Mapping, context value alignment to JIRA_ISSUE_KEY, and automatic fallback retry mechanism in the following job:
Google Chronicle: Version 92.0
Updated the action to use new search API in the following action:
CrowdStrike Falcon: Version 80.0
Implemented 500-device safety limit per entity search, updated device sorting by last_seen.desc, and added batch request chunking for device details, login history, and online states in the following action:
From August 3 through August 5, 2026, the following features will be automatically enabled for Looker (Google Cloud core) instances running Looker 26.12.
Looker Continuous Integration (CI) now supports email alerts. When you create or edit a CI suite, you can enable the Enable email alerts toggle to specify email recipients and select which run statuses will trigger emails (Failed, Error, Passed, or Cancelled). For more information, see Set up alerting.
The custom calendar feature is now generally available.
The Expression Assistant is now generally available.
Localization is now supported for the dimension_group parameter. You can localize the timeframes, intervals, or custom timeframes generated by a dimension group by providing translations in your locale strings files. For more details, see Localizing dimension groups.
The LookML Projects page has been updated with a more performant tabbed layout, which features three tabs: Models and Projects, Pending Projects, and Marketplace Projects.
Looker admins now have the ability to configure a Looker instance to require multi-factor authentication (MFA) whenever a user tries to log in by using an email and a password. This feature is enabled by default.
The Enhanced search feature is now generally available.
Now available in preview, the Admin Assistant helps you use natural language to manage Looker roles.
Now available in preview, enhanced observability metrics — including engagement and estimated token usage data — are available for Conversational Analytics on the Conversational Analytics System Activity dashboard. To enable this feature, a Looker admin must turn on the Conversational Analytics Agent Token usage setting on the General page in the Preview section of the Admin panel.
Now available in preview, the new Modern User Interface feature enables modernized layouts and design alongside new configuration settings for visualizations and dashboards. When this preview feature is enabled, users can apply a Modern visualization theme that features updated typography and modern, accessible color palettes for improved data legibility. Additionally, a new Modern dashboard style provides a high-density, streamlined design that optimizes data viewing and aligns with Google's latest design standards.
The Insight Assistant now displays the process the assistant uses to generate the response, showing key details in your data that it used to generate the response, and listing the fields from your Explore that it used.
When you chat in Gemini Enterprise with data agents that you create in Looker, agent responses now include charts and visualizations.
Complimentary Data Studio Pro licenses aren't available for Looker (Google Cloud core) instances that are created after August 1, 2026.
(Airflow 3.2.2, 3.1.8, and 2.11.1)
The apache-airflow-providers-google package was upgraded to version 22.2.2.
For more information about changes, see the
apache-airflow-providers-google changelog.
New Airflow builds are available in Managed Airflow (Gen 3):
These builds are versions with an extended upgrade timeline.
New images are available in Managed Airflow (Gen 2):
These images are versions with an extended upgrade timeline.
The following Managed Airflow versions and builds have reached their end of support period: composer-3-airflow-2.10.5-build.11, composer-3-airflow-2.9.3-build.31, composer-2.13.9-airflow-2.9.3, and composer-2.13.9-airflow-2.10.5.
]]>AlloyDB now supports Best Matching 25 (BM25) indexes for full-text search in
Preview.
You can use the pg_textsearch extension to create BM25 indexes and optimize
probabilistic ranking of full-text search. This feature is supported on AlloyDB
instances running PostgreSQL 17 or 18.
For more information, see Create and manage a BM25 index.
Cloud CDN supports native image optimization at the Google network edge for global external Application Load Balancers. This feature offloads compute-intensive image transformations, such as resizing, cropping, and format conversion to reduce origin server load and egress costs. This feature is in Preview.
For more information, see Optimize images with Cloud CDN.
Regular expression URL rewrites (regexRewrite) for route rules in URL maps are
now available for Application Load Balancers. You can use regular expression
pattern rewrite actions to rewrite URL paths by substituting or removing URL
path components before forwarding requests to your backends.
For more information, see Regular expression URL rewrites for route rules.
This feature is in Preview.
DNS automation is now generally available (GA) on Cloud SQL instances where Private Service Connect is enabled.
You can use DNS automation to provision and manage per-instance DNS records automatically. On Enterprise Plus edition instances where DNS automation is enabled, you can also enable a global write endpoint DNS that automatically resolves to your current primary instance.
Cloud SQL for MySQL supports resource groups. MySQL resource groups let you manage resource allocation for different workloads on your Cloud SQL for MySQL instance. By using resource groups, you can prevent less important workloads from consuming excessive CPU or memory resources.
To use MySQL resource groups, you must have maintenance version MYSQL_VERSION.R20260320.00_20 or later installed on your instance.
For more information, see Manage CPU allocation with MySQL resource groups.
DNS automation is now generally available (GA) on Cloud SQL instances where Private Service Connect is enabled.
You can use DNS automation to provision and manage per-instance DNS records automatically. On Enterprise Plus edition instances where DNS automation is enabled, you can also enable a global write endpoint DNS that automatically resolves to your current primary instance.
DNS automation is now generally available (GA) on Cloud SQL instances where Private Service Connect is enabled.
You can use DNS automation to provision and manage per-instance DNS records automatically. On Enterprise Plus edition instances where DNS automation is enabled, you can also enable a global write endpoint DNS that automatically resolves to your current primary instance.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.18.39 | v29.4.3 | v2.3.2 | See List |
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.12.94 | v27.5.1 | v2.2.6 | See List |
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.18.41 | v29.4.3 | v2.3.2 | See List |
/dev/hugepages is now mounted with the noexec option.
Updated containerd to v2.2.6.
Added a splash screen which is displayed at login.
Added a splash screen which is displayed at login.
cchost: Add bpf-lsm-policy for VM restrictions.
Added support for net-fs/lustre-client-drivers v2.14.0_p256.
Added support for net-fs/lustre-client-drivers v2.14.0_p256.
Fixed CVE-2026-39827, CVE-2026-39828, CVE-2026-39829, CVE-2026-39830, CVE-2026-39831, CVE-2026-39832, CVE-2026-39833, CVE-2026-39834, CVE-2026-39835, CVE-2026-42508, CVE-2026-46595, CVE-2026-46597, and CVE-2026-46598 in dev-go/crypto.
Added support for the swiotlb=any kernel command line parameter.
Added support for the Lustre 2.14.0_p246 drivers.
Fixed CVE-2026-58470 in net-misc/wget.
Updated google-guest-configs to v20260121.00.
Added support for the Lustre 2.14.0_p249 drivers.
Fixed CVE-2026-59890 in dev-python/setuptools.
Added support for the R595 Nvidia driver production branch.
Fixed CVE-2026-64244 in the Linux kernel.
Updated the Linux kernel to v6.18.41.
Added support for the swiotlb=any kernel command line parameter.
Fixed CVE-2026-64247 in the Linux kernel.
cchost: Add bpf-lsm-policy for VM restrictions.
Allow overriding IMA policy from oem partition.
Fixed CVE-2026-64253 in the Linux kernel.
cchost: Increased the size of the kernel partitions from 16 MiB to 32 MiB.
Apply hardening sysctls on cchost boards.
Fixed CVE-2026-64265 in the Linux kernel.
Dropped support for the NVIDIA 535 drivers.
Fixed CVE-2026-64266 in the Linux kernel.
Added TPUDirect support.
Enabled mm hardening kernel cmdlines on cchost.
Fixed CVE-2026-64284 in the Linux kernel.
Added support for zswap in the Linux kernel.
Fixed the "CrackArmor" vulnerability in the Linux kernel.
Fixed CVE-2026-64289 in the Linux kernel.
Enabled CONFIG_MEMORY_FAILURE in the Linux kernel for ARM64. This should improve memory errors handling when running CUDA workloads.
Fixes a kernel panic in virtio_pci teardown when virtually queues are conditionally skipped.
Fixed CVE-2026-64294 in the Linux kernel.
Added support for NVIDIA GRID driver version 580.159.03.
Increased the size of the EFI partition from 32 MiB to 64 MiB and increased the sizes of both kernel partitions from 16 MiB to 32 MiB on x86.
Fixed CVE-2026-64298 in the Linux kernel.
Changed google-guest-agent's plugin installation path to /var/lib/google/guest-agent.
Made it so that /etc/machine-id is mounted with noexec, nosuid, and nodev.
Fixed CVE-2026-64299 in the Linux kernel.
Update udev rule for protected_stateful_partition
Made it so that /run is mounted as noexec.
Fixed CVE-2026-64306 in the Linux kernel.
Updated containerd to v2.3.2., containerd-test to v2.3.2.. This resolves CVE-2026-46680, CVE-2026-50195, CVE-2026-53488, CVE-2026-53492.
On cchost boards, autoload IMA policy on boot.
Fixed CVE-2026-64313 in the Linux kernel.
Updated dev-cpp/abseil-cpp to v20230802.0; dev-libs/flatbuffersto v24.3.25; sys-devel/autofdo to v0.30-r12; media-libs/cros-camera-hal-fake to v0.0.1-r616; chromeos-base/chromeos-dbus-bindings to v0.0.1-r2800; chromeos-base/chromeos-installer to v0.1.0-r4432; chromeos-base/hardware_verifier_proto to v0.0.1-r819; chromeos-base/imageloader to v0.0.1-r2064; chromeos-base/libbrillo to v0.0.1-r2547; chromeos-base/libchrome to v0.0.1-r1242; chromeos-base/libhwsec-foundation to v0.0.1-r826; chromeos-base/libstorage to v0.0.1-r162; chromeos-base/metrics to v0.0.2-r3890; chromeos-base/perfetto to v48.1-r69; chromeos-base/quipper to v0.0.1-r3050; chromeos-base/system_api to v0.0.1-r5972; chromeos-base/update_engine to v0.0.3-r5184; chromeos-base/vboot_reference to v1.0-r2986;dev-rust/vboot_reference-sys to v1.0.0-r36; chromeos-base/verity to v0.0.1-r583; chromeos-base/vpd to v0.0.1-r339; Removed packages dev-util/bazel,dev-util/iwyu, and dev-util/cvise; Removed chromeos-base/crash-reporter.
Set static UUID for the stateful partition.
Fixed CVE-2026-64317 in the Linux kernel.
Updated dev-lang/rust, dev-lang/rust-host, dev-lang/rust-bootstrap to v1.84.1; dev-rust/protobuf-codegen to v2.28.0; dev-rust/system_api to v0.24.53-r1596; dev-rust/third-party-crates-src to v0.0.1-r288.
Switch cchost-* boards to legacy iptables.
Fixed CVE-2026-64319 in the Linux kernel.
Updated dev-libs/isa-l to v2.32.1.
Update sys-process/audit to v3.0.9.
Fixed CVE-2026-64320 in the Linux kernel.
Updated sys-devel/gdb to v15.1; sys-apps/flashrom to v0.9.9-r1758; dev-python/cryptography to v43.0.3; Update dev-python/pyopenssl to v24.2.1.
Updated app-arch/gzip to v1.14_p20260502, app-arch/xz-utils to v5.4.7, app-arch/zstd to v1.5.7, app-crypt/mit-krb5 to v1.22.2, app-editors/vim to v9.1.2148, app-editors/vim-core to v9.1.2148, dev-libs/libaio to v0.3.113-r2, dev-libs/xxhash to v0.8.3, dev-python/PySocks to v1.6.8, dev-python/markupsafe to v2.1.5, dev-python/pyrsistent to v0.14.11, dev-python/python-magic to v0.4.27, dev-python/pyyaml to v6.0.3, dev-python/rfc3339-validator to v0.1.4-r1, net-misc/rsync to v3.4.4, sys-apps/mawk to v1.3.4_p20260302, sys-libs/libxcrypt to v4.4.38, sys-libs/talloc to v2.4.4, sys-fs/lvm2 to v2.03.39, sys-libs/binutils-libs to v2.46.1.. This resolves CVE-2025-69644.
Fixed CVE-2026-64322 in the Linux kernel.
Updated sys-devel/llvm to v20.0_pre547379; sys-libs/compiler-rt to v20.0_pre547379; sys-libs/libcxx to v20.0_pre547379; sys-libs/llvm-libunwind to v20.0_pre547379.
Updated app-containers/cloud-provider-gcp to v35.0.8, app-containers/docker-credential-gcr to v2.1.32, app-containers/nvidia-container-toolkit to v1.17.9, dev-python/oauthlib to v3.0.2, net-fs/nfs-utils to v2.6.4, sys-libs/libapparmor to v3.1.7..
Fixed CVE-2026-64323 in the Linux kernel.
Upgrade Kubernetes to 1.36.1
Fixed CVE-2026-64324 in the Linux kernel.
Upgraded app-admin/oslogin to v20260626.00.
Updated dev-libs/libtraceevent to v1.7.3, dev-libs/libtracefs to v1.6.4, net-firewall/conntrack-tools to v1.4.9, sys-fs/fuse to v2.9.9, sys-fs/fuse-common to v3.10.5.
Fixed CVE-2026-64326 in the Linux kernel.
Upgraded app-crypt/sbsigntools to v0.42.0, and net-misc/m2crypto to v0.42.0.
Fixed CVE-2026-64354 in the Linux kernel.
Upgraded chromeos-base/debugd-client to v0.0.1-r2739.
Updated google-guest-configs to v20260121.00.
Fixed CVE-2026-64355 in the Linux kernel.
Upgraded dev-db/sqlite to v3.53.3.
Updated sys-apps/casfs to v0.1.14.
Fixed CVE-2026-64357 in the Linux kernel.
Upgraded dev-libs/expat to v2.8.2.
Updated sys-libs/pam to v1.5.3.
Fixed CVE-2026-64368 in the Linux kernel.
Upgraded dev-libs/openssl from v3.5.6 to v4.0.0.
Fixed CVE-2026-64370 in the Linux kernel.
Upgraded net-misc/chrony to v4.8-r2.
Fixed CVE-2026-64373 in the Linux kernel.
Upgraded net-misc/curl to 8.21.0.
Fixed CVE-2026-64378 in the Linux kernel.
Upgraded net-misc/socat to v1.8.1.3.
Fixed CVE-2026-64379 in the Linux kernel.
Upgraded sys-apps/acl to v2.4.0.
Fixed CVE-2026-64380 in the Linux kernel.
Upgraded sys-apps/attr to v2.6.0.
Fixed CVE-2026-64381 in the Linux kernel.
Updated the Linux kernel to v6.18.39.
Fixed CVE-2026-64382 in the Linux kernel.
Fixed CVE-2026-29111 in sys-apps/systemd
Updated uhaul to v6.18-0.
Fixed CVE-2026-64383 in the Linux kernel.
Fixed CVE-2026-39827, CVE-2026-39828, CVE-2026-39829, CVE-2026-39830, CVE-2026-39831, CVE-2026-39832, CVE-2026-39833, CVE-2026-39834, CVE-2026-39835, CVE-2026-42508, CVE-2026-46595, CVE-2026-46597, and CVE-2026-46598 in dev-go/crypto.
Fixed CVE-2026-64384 in the Linux kernel.
Fixed CVE-2026-40225 in sys-apps/systemd.
Fixed CVE-2026-64385 in the Linux kernel.
Fixed CVE-2026-40355 and CVE-2026-40356 in app-crypt/mit-krb5.
Fixed CVE-2026-64386 in the Linux kernel.
Fixed CVE-2026-58469, CVE-2026-58471, CVE-2026-58472 in net-misc/wget.
Fixed CVE-2026-64387 in the Linux kernel.
Fixed CVE-2026-58470 in net-misc/wget.
Upgraded sys-apps/ek-cpu-balloon to v1.2.3.
Fixed CVE-2026-64411 in the Linux kernel.
Fixed CVE-2026-59890 in dev-python/setuptools.
Upgraded sys-apps/iproute2 to version 6.18.0.
Fixed CVE-2026-64412 in the Linux kernel.
Upgraded sys-apps/xemu to v0.0.9.
Fixed CVE-2026-64414 in the Linux kernel.
Updated glib and gdbus-codegen to v2.89.1. This resolves CVE-2026-58013,CVE-2026-58014,CVE-2026-58015,CVE-2026-58016.
Upgraded sys-fs/cryptsetup to v2.8.6.
Fixed CVE-2026-64415 in the Linux kernel.
Upgraded net-misc/openssh to version 10.3_p1. This resolves CVE-2026-35387 and CVE-2026-35388.
Upgraded sysram to v6.18-0.
Fixed CVE-2026-64418 in the Linux kernel.
Runtime sysctl changes:
cchost: Add bpf-lsm-policy for VM restrictions.
Fixed CVE-2026-64422 in the Linux kernel.
cchost: Increased the size of the kernel partitions from 16 MiB to 32 MiB.
Fixed CVE-2026-64423 in the Linux kernel.
Fixed CVE-2026-64425 in the Linux kernel.
Added TPUDirect support.
Fixed CVE-2026-64432 in the Linux kernel.
Added nvidia-fs support to the COS GPU installer.
Fixed CVE-2026-64435 in the Linux kernel.
Added support for 590.44.01 and 590.48.01 NVIDIA driver for NVIDIA_RTX_PRO_6000
Fixed CVE-2026-64436 in the Linux kernel.
Added support for 8th generation TPU devices.
Fixed CVE-2026-64448 in the Linux kernel.
Added support for NVIDIA driver v535.288.01, v570.211.01 and v580.126.09.
Fixed CVE-2026-64456 in the Linux kernel.
Added support for larger ring sizes for the GVNIC driver in DQO-QPL mode.
Fixed CVE-2026-64473 in the Linux kernel.
Added support for loading the ublk kernel module.
Fixed CVE-2026-64474 in the Linux kernel.
Added support for zswap in the Linux kernel.
Fixed CVE-2026-64475 in the Linux kernel.
Added the cos_kernel_args tool that allows manipulating kernel command line arguments of a COS image.
Fixed CVE-2026-64512 in the Linux kernel.
Changed default sysctl networking values on A4x-max machine type only.
Fixed CVE-2026-64514 in the Linux kernel.
Enabled CONFIG_MEMORY_FAILURE in the Linux kernel for ARM64. This should improve memory errors handling when running CUDA workloads.
Fixed CVE-2026-64556 in the Linux kernel.
Enabled dynamic configuration of FUSE max pages limit.
Upgraded net-libs/nghttp2 to 1.69.0 and fixed CVE-2026-58055.
Enabled dynamic debug in the Linux kernel.
Runtime sysctl changes:
Reverted iproute2 to v5.16.0.
Switched to using systemd-resolved stub resolver by default, which fixes DNS caching issues.
Added support for NVIDIA GRID driver version 580.159.03.
Added support for NVIDIA driver v580.126.09-grid for NVIDIA_RTX_PRO_6000 GPU type.
Added support for NVIDIA driver v580.159.03.
Added support for NVIDIA driver v580.159.04.
Added support for NVIDIA driver v595.71.05.
Added support for NVIDIA drivers v580.126.16 and v580.126.20.
Changed google-guest-agent's plugin installation path to /var/lib/google/guest-agent.
Dropped support for NVIDIA MFT Tools v4.32.0.
Enabled buffer overflow detection for kernel str/mem functions.
Fixed a crash that occurs when using the configfile or
source GRUB2 commands when Secure Boot is enabled.
Fixed a kernel bug which could cause traffic drops after NIC resets.
Fixed an ek-cpu-balloon bug which would result in CPUs being underreported on ek machines with SMT enabled.
Update udev rule for protected_stateful_partition
Updated containerd to v2.3.2., containerd-test to v2.3.2.. This resolves CVE-2026-35469, CVE-2026-46680, CVE-2026-50195, CVE-2026-53488, CVE-2026-53492.
Updated dev-cpp/abseil-cpp to v20230802.0; dev-libs/flatbuffersto v24.3.25; sys-devel/autofdo to v0.30-r12; media-libs/cros-camera-hal-fake to v0.0.1-r616; chromeos-base/chromeos-dbus-bindings to v0.0.1-r2800; chromeos-base/chromeos-installer to v0.1.0-r4432; chromeos-base/hardware_verifier_proto to v0.0.1-r819; chromeos-base/imageloader to v0.0.1-r2064; chromeos-base/libbrillo to v0.0.1-r2547; chromeos-base/libchrome to v0.0.1-r1242; chromeos-base/libhwsec-foundation to v0.0.1-r826; chromeos-base/libstorage to v0.0.1-r162; chromeos-base/metrics to v0.0.2-r3890; chromeos-base/perfetto to v48.1-r69; chromeos-base/quipper to v0.0.1-r3050; chromeos-base/system_api to v0.0.1-r5972; chromeos-base/update_engine to v0.0.3-r5184; chromeos-base/vboot_reference to v1.0-r2986;dev-rust/vboot_reference-sys to v1.0.0-r36; chromeos-base/verity to v0.0.1-r583; chromeos-base/vpd to v0.0.1-r339; Removed packages dev-util/bazel,dev-util/iwyu, and dev-util/cvise; Removed chromeos-base/crash-reporter.
Updated dev-lang/rust, dev-lang/rust-host, dev-lang/rust-bootstrap to v1.84.1; dev-rust/protobuf-codegen to v2.28.0; dev-rust/system_api to v0.24.53-r1596; dev-rust/third-party-crates-src to v0.0.1-r288.
Updated dev-libs/isa-l to v2.32.1.
Updated sys-devel/gdb to v15.1; sys-apps/flashrom to v0.9.9-r1758; dev-python/cryptography to v43.0.3; Update dev-python/pyopenssl to v24.2.1.
Updated sys-devel/llvm to v20.0_pre547379; sys-libs/compiler-rt to v20.0_pre547379; sys-libs/libcxx to v20.0_pre547379; sys-libs/llvm-libunwind to v20.0_pre547379.
Upgrade Kubernetes to 1.36.1
Upgraded CASFS to v0.1.3.
Upgraded app-admin/google-guest-agent to v20260121.00.
Upgraded app-admin/google-osconfig-agent to v20260119.00.
Upgraded app-admin/logrotate to v3.22.0-r1.
Upgraded app-admin/oslogin to v20260626.00.
Upgraded app-admin/sosreport to v4.11.2.
Upgraded app-arch/unzip to v6.0_p29-r2.
Upgraded app-containers/cni-plugins to v1.9.1.
Upgraded app-containers/docker to v29.4.3, Upgraded app-containers/docker-test to v29.4.3, Upgraded app-containers/docker-cli to v29.4.3.
Upgraded app-containers/docker-credential-helpers to v0.9.8.
Upgraded app-crypt/sbsigntools to v0.42.0, and net-misc/m2crypto to v0.42.0.
Upgraded app-emulation/cloud-init to v26.1.
Upgraded app-shells/dash to v0.5.13.4-r2.
Upgraded chromeos-base/chromeos-common-script to v0.0.1-r672.
Upgraded chromeos-base/debugd-client to v0.0.1-r2739.
Upgraded chromeos-base/google-breakpad to v2026.06.22.165940-r278.
Upgraded chromeos-base/power_manager-client to v0.0.1-r2973.
Upgraded chromeos-base/session_manager-client to v0.0.1-r2834.
Upgraded cos-gpu-installer to v2.7.4.
Upgraded dev-db/sqlite to v3.53.3.
Upgraded dev-libs/expat to v2.8.2.
Upgraded dev-libs/openssl from v3.5.6 to v4.0.0.
Upgraded dev-util/gn to v2331.
Upgraded dev-utils/gdbus-codegen to v2.86.3.
Upgraded net-firewall/iptables to v1.8.13.
Upgraded net-libs/libnetfilter_conntrack to v1.1.1.
Upgraded net-libs/libnetfilter_queue to v1.0.5-r1.
Upgraded net-misc/chrony to v4.8-r2.
Upgraded net-misc/curl to 8.21.0. to fix CVE-2025-13034, CVE-2025-14017, CVE-2025-14524, CVE-2025-14819, CVE-2025-15079, CVE-2025-15224, CVE-2026-1965, CVE-2026-3783, CVE-2026-4873, CVE-2026-5545, CVE-2026-5773, CVE-2026-6253, CVE-2026-6276, CVE-2026-6429, CVE-2026-7009, CVE-2026-7168.
Upgraded net-misc/socat to v1.8.1.3.
Upgraded sys-apps/acl to v2.4.0.
Upgraded sys-apps/attr to v2.6.0.
Upgraded sys-apps/file to v5.47-r1.
Upgraded sys-apps/gentoo-functions to v1.7.7.
Upgraded sys-apps/hwdata to v0.401.
Upgraded sys-apps/less to v704.
Upgraded sys-apps/makedumpfile to v1.7.9.
Upgraded sys-apps/pv to v1.10.4.
Upgraded sys-libs/libcap to v2.78.
Upgraded sys-libs/libcap-ng to v0.9.3.
Upgraded sys-libs/zlib to v1.3.2-r1.
Upgraded sys-process/lsof to v4.99.6.
Upgraded sys-process/procps to v4.0.6.
Upgraded the dump capture kernel to Linux v6.18.
Upgraded the galog version to v0.0.0-20250924170816-9dbf105986f4 in google-guest-agent to fix an issue with high CPU consumption.
Upgraded virtual/logger to v0-r3.
Fixed CVE-2025-15281 and CVE-2026-0861 in sys-libs/glibc.
Fixed CVE-2025-40147 in the Linux kernel.
Fixed CVE-2026-0915 in sys-apps/glibc.
Fixed CVE-2026-0994 in dev-libs/protobuf.
Fixed CVE-2026-27135 in net-libs/nghttp2.
Fixed CVE-2026-29111 in sys-apps/systemd
Fixed CVE-2026-32597 with pyjwt package upgrade to v2.12.1.
Fixed CVE-2026-33997 and CVE-2026-34040 in Docker.
Fixed CVE-2026-34743 in app-arch/xz-utils.
Fixed CVE-2026-35385 and CVE-2026-35386 in net-misc/openssh.
Fixed CVE-2026-35414 in net-misc/openssh.
Fixed CVE-2026-40225 in sys-apps/systemd.
Fixed CVE-2026-40226 in sys-apps/systemd.
Fixed CVE-2026-40355 and CVE-2026-40356 in app-crypt/mit-krb5.
Fixed CVE-2026-4046 in sys-libs/glibc.
Fixed CVE-2026-4437,CVE-2026-4438 in sys-libs/glibc.
Fixed CVE-2026-44431 in dev-python/urllib3.
Fixed CVE-2026-50195, CVE-2026-53488, CVE-2026-53492, CVE-2026-53489, and CVE-2026-47262 in app-containers/containerd.
Fixed CVE-2026-58469, CVE-2026-58471, CVE-2026-58472 in net-misc/wget.
Fixed CVE-2026-5928 in sys-libs/glibc.
Fixed CVE-2026-6238 in sys-libs/glibc.
Fixed CVE-2026-6732 in dev-libs/libxml2.
Fixed CVE-2026-7210 in dev-lang/python.
Fixed EFI variable OOB read in grub config parsing.
Fixed KCTF-329f0b9 in the Linux kernel.
Fixed KCTF-7cb9a23 in the Linux kernel.
Fixed KCTF-c9bc175 in the Linux kernel.
Fixed KCTF-e3f000f in the Linux kernel.
Fixed KCTF-f8db647 in the Linux kernel.
Fixed argument injection in toolbox.
Updated dev-lang/go to 1.25.10. This fixes CVE-2026-33814,CVE-2026-39819,CVE-2026-39823,CVE-2026-39825,CVE-2026-42499,CVE-2026-39817,CVE-2026-39820,CVE-2026-39826,CVE-2026-39836.
Updated dev-libs/libxml2 to version 2.14.6. This resolves CVE-2025-6021.
Updated dev-python/pyjwt to v2.13.0. This fixes CVE-2026-48522, CVE-2026-48524, CVE-2026-48525, CVE-2026-485256.
Updated glib to v2.89.1., gdbus-codegen to v2.89.1.. This resolves CVE-2026-58013, CVE-2026-58014, CVE-2026-58015, CVE-2026-58016.
Updated go to v1.25.9. This resolves CVE-2026-32280, CVE-2026-32281, CVE-2026-32283, CVE-2026-27140, CVE-2026-27144.
Upgraded dev-libs/glib to v2.86.3. This fixes CVE-2025-14087, CVE-2025-14512 and CVE-2025-13601.
Upgraded dev-libs/libgcrypt to v1.10.4 to fix CVE-2026-41989.
Upgraded dev-libs/openssl to v3.5.7 to fix CVE-2025-15467, CVE-2026-28387, CVE-2026-28388, CVE-2026-28389, CVE-2026-28390, CVE-2026-31790, CVE-2026-34180, CVE-2026-34181, CVE-2026-34182, CVE-2026-34183, CVE-2026-42764, CVE-2026-45445, CVE-2026-45447, CVE-2026-7383, CVE-2026-9076.
Upgraded net-misc/openssh to 10.3_p1. to fix CVE-2026-35387, CVE-2026-35388.
Runtime sysctl changes:
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.12.94 | v27.5.1 | v2.1.9 | See List |
cchost: Add bpf-lsm-policy for VM restrictions.
Added kernel patch to reduce bcache garbage collection sleep interval to prevent I/O stalls.
Fixed CVE-2026-35177 in app-editors/vim and app-editors/vim-core.
Fixed CVE-2026-39827, CVE-2026-39828, CVE-2026-39829, CVE-2026-39830, CVE-2026-39831, CVE-2026-39832, CVE-2026-39833, CVE-2026-39834, CVE-2026-39835, CVE-2026-42508, CVE-2026-46595, CVE-2026-46597, and CVE-2026-46598 in dev-go/crypto.
Fixed CVE-2026-58470 in net-misc/wget.
Fixed CVE-2026-59890 in dev-python/setuptools.
Fixed CVE-2026-64227 in the Linux kernel.
Fixed CVE-2026-64244 in the Linux kernel.
Fixed CVE-2026-64247 in the Linux kernel.
Fixed CVE-2026-64265 in the Linux kernel.
Fixed CVE-2026-64266 in the Linux kernel.
Fixed CVE-2026-64284 in the Linux kernel.
Fixed CVE-2026-64289 in the Linux kernel.
Fixed CVE-2026-64294 in the Linux kernel.
Fixed CVE-2026-64298 in the Linux kernel.
Fixed CVE-2026-64299 in the Linux kernel.
Fixed CVE-2026-64306 in the Linux kernel.
Fixed CVE-2026-64313 in the Linux kernel.
Fixed CVE-2026-64317 in the Linux kernel.
Fixed CVE-2026-64319 in the Linux kernel.
Fixed CVE-2026-64322 in the Linux kernel.
Fixed CVE-2026-64323 in the Linux kernel.
Fixed CVE-2026-64324 in the Linux kernel.
Fixed CVE-2026-64326 in the Linux kernel.
Fixed CVE-2026-64354 in the Linux kernel.
Fixed CVE-2026-64357 in the Linux kernel.
Fixed CVE-2026-64368 in the Linux kernel.
Fixed CVE-2026-64370 in the Linux kernel.
Fixed CVE-2026-64373 in the Linux kernel.
Fixed CVE-2026-64378 in the Linux kernel.
Fixed CVE-2026-64379 in the Linux kernel.
Fixed CVE-2026-64381 in the Linux kernel.
Fixed CVE-2026-64382 in the Linux kernel.
Fixed CVE-2026-64383 in the Linux kernel.
Fixed CVE-2026-64384 in the Linux kernel.
Fixed CVE-2026-64385 in the Linux kernel.
Fixed CVE-2026-64386 in the Linux kernel.
Fixed CVE-2026-64387 in the Linux kernel.
Fixed CVE-2026-64411 in the Linux kernel.
Fixed CVE-2026-64412 in the Linux kernel.
Fixed CVE-2026-64414 in the Linux kernel.
Fixed CVE-2026-64415 in the Linux kernel.
Fixed CVE-2026-64418 in the Linux kernel.
Fixed CVE-2026-64422 in the Linux kernel.
Fixed CVE-2026-64423 in the Linux kernel.
Fixed CVE-2026-64425 in the Linux kernel.
Fixed CVE-2026-64432 in the Linux kernel.
Fixed CVE-2026-64435 in the Linux kernel.
Fixed CVE-2026-64436 in the Linux kernel.
Fixed CVE-2026-64448 in the Linux kernel.
Fixed CVE-2026-64456 in the Linux kernel.
Fixed CVE-2026-64473 in the Linux kernel.
Fixed CVE-2026-64474 in the Linux kernel.
Fixed CVE-2026-64475 in the Linux kernel.
Fixed CVE-2026-64512 in the Linux kernel.
Fixed CVE-2026-64514 in the Linux kernel.
Upgraded net-libs/nghttp2 to 1.69.0 and fixed CVE-2026-58055.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.6.143 | v27.5.1 | v2.0.10 | See List |
Fixed CVE-2026-39827, CVE-2026-39828, CVE-2026-39829, CVE-2026-39830, CVE-2026-39831, CVE-2026-39832, CVE-2026-39833, CVE-2026-39834, CVE-2026-39835, CVE-2026-42508, CVE-2026-46595, CVE-2026-46597, and CVE-2026-46598 in dev-go/crypto.
Fixed CVE-2026-58470 in net-misc/wget.
Fixed CVE-2026-59890 in dev-python/setuptools.
Fixed CVE-2026-64227 in the Linux kernel.
Fixed CVE-2026-64266 in the Linux kernel.
Fixed CVE-2026-64286 in the Linux kernel.
Fixed CVE-2026-64287 in the Linux kernel.
Fixed CVE-2026-64294 in the Linux kernel.
Fixed CVE-2026-64298 in the Linux kernel.
Fixed CVE-2026-64299 in the Linux kernel.
Fixed CVE-2026-64306 in the Linux kernel.
Fixed CVE-2026-64313 in the Linux kernel.
Fixed CVE-2026-64317 in the Linux kernel.
Fixed CVE-2026-64322 in the Linux kernel.
Fixed CVE-2026-64323 in the Linux kernel.
Fixed CVE-2026-64324 in the Linux kernel.
Fixed CVE-2026-64326 in the Linux kernel.
Fixed CVE-2026-64355 in the Linux kernel.
Fixed CVE-2026-64368 in the Linux kernel.
Fixed CVE-2026-64370 in the Linux kernel.
Fixed CVE-2026-64373 in the Linux kernel.
Fixed CVE-2026-64379 in the Linux kernel.
Fixed CVE-2026-64380 in the Linux kernel.
Fixed CVE-2026-64381 in the Linux kernel.
Fixed CVE-2026-64382 in the Linux kernel.
Fixed CVE-2026-64383 in the Linux kernel.
Fixed CVE-2026-64384 in the Linux kernel.
Fixed CVE-2026-64385 in the Linux kernel.
Fixed CVE-2026-64386 in the Linux kernel.
Fixed CVE-2026-64387 in the Linux kernel.
Fixed CVE-2026-64411 in the Linux kernel.
Fixed CVE-2026-64412 in the Linux kernel.
Fixed CVE-2026-64422 in the Linux kernel.
Fixed CVE-2026-64423 in the Linux kernel.
Fixed CVE-2026-64425 in the Linux kernel.
Fixed CVE-2026-64435 in the Linux kernel.
Fixed CVE-2026-64436 in the Linux kernel.
Fixed CVE-2026-64448 in the Linux kernel.
Fixed CVE-2026-64456 in the Linux kernel.
Fixed CVE-2026-64475 in the Linux kernel.
Fixed CVE-2026-64507 in the Linux kernel.
Fixed CVE-2026-64508 in the Linux kernel.
Fixed CVE-2026-64512 in the Linux kernel.
Fixed CVE-2026-64514 in the Linux kernel.
Fixed CVE-2026-64530 in the Linux kernel.
Fixed CVE-2026-64534 in the Linux kernel.
Fixed CVE-2026-64538 in the Linux kernel.
Fixed CVE-2026-64545 in the Linux kernel.
Fixed CVE-2026-64546 in the Linux kernel.
Fixed CVE-2026-64548 in the Linux kernel.
Fixed CVE-2026-64552 in the Linux kernel.
Fixed CVE-2026-64554 in the Linux kernel.
Fixed CVE-2026-64556 in the Linux kernel.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.6.143 | v24.0.9 | v1.7.34 | See List |
Fixed CVE-2026-35177 in app-editors/vim and app-editors/vim-core.
Fixed CVE-2026-39827, CVE-2026-39828, CVE-2026-39829, CVE-2026-39830, CVE-2026-39831, CVE-2026-39832, CVE-2026-39833, CVE-2026-39834, CVE-2026-39835, CVE-2026-42508, CVE-2026-46595, CVE-2026-46597, and CVE-2026-46598 in dev-go/crypto.
Fixed CVE-2026-58470 in net-misc/wget.
Fixed CVE-2026-59890 in dev-python/setuptools.
Fixed CVE-2026-64244 in the Linux kernel.
Fixed CVE-2026-64247 in the Linux kernel.
Fixed CVE-2026-64266 in the Linux kernel.
Fixed CVE-2026-64294 in the Linux kernel.
Fixed CVE-2026-64298 in the Linux kernel.
Fixed CVE-2026-64299 in the Linux kernel.
Fixed CVE-2026-64306 in the Linux kernel.
Fixed CVE-2026-64313 in the Linux kernel.
Fixed CVE-2026-64317 in the Linux kernel.
Fixed CVE-2026-64322 in the Linux kernel.
Fixed CVE-2026-64323 in the Linux kernel.
Fixed CVE-2026-64324 in the Linux kernel.
Fixed CVE-2026-64326 in the Linux kernel.
Fixed CVE-2026-64355 in the Linux kernel.
Fixed CVE-2026-64368 in the Linux kernel.
Fixed CVE-2026-64370 in the Linux kernel.
Fixed CVE-2026-64373 in the Linux kernel.
Fixed CVE-2026-64379 in the Linux kernel.
Fixed CVE-2026-64380 in the Linux kernel.
Fixed CVE-2026-64381 in the Linux kernel.
Fixed CVE-2026-64382 in the Linux kernel.
Fixed CVE-2026-64383 in the Linux kernel.
Fixed CVE-2026-64384 in the Linux kernel.
Fixed CVE-2026-64385 in the Linux kernel.
Fixed CVE-2026-64386 in the Linux kernel.
Fixed CVE-2026-64387 in the Linux kernel.
Fixed CVE-2026-64411 in the Linux kernel.
Fixed CVE-2026-64412 in the Linux kernel.
Fixed CVE-2026-64422 in the Linux kernel.
Fixed CVE-2026-64423 in the Linux kernel.
Fixed CVE-2026-64425 in the Linux kernel.
Fixed CVE-2026-64435 in the Linux kernel.
Fixed CVE-2026-64436 in the Linux kernel.
Fixed CVE-2026-64448 in the Linux kernel.
Fixed CVE-2026-64456 in the Linux kernel.
Fixed CVE-2026-64474 in the Linux kernel.
Fixed CVE-2026-64475 in the Linux kernel.
Fixed CVE-2026-64507 in the Linux kernel.
Fixed CVE-2026-64508 in the Linux kernel.
Fixed CVE-2026-64512 in the Linux kernel.
Fixed CVE-2026-64514 in the Linux kernel.
Fixed CVE-2026-64530 in the Linux kernel.
Fixed CVE-2026-64534 in the Linux kernel.
Fixed CVE-2026-64538 in the Linux kernel.
Fixed CVE-2026-64545 in the Linux kernel.
Fixed CVE-2026-64546 in the Linux kernel.
Fixed CVE-2026-64552 in the Linux kernel.
Fixed CVE-2026-64554 in the Linux kernel.
Upgraded net-libs/nghttp2 to 1.69.0 and fixed CVE-2026-58055.
Gemini Enterprise: End-to-end tracing support for data connectors
Tracing support is extended end-to-end across the data connector workflow, introducing two new trace spans for better observability:
execute_tool: Represents the execution of a tool on the agent
orchestration layer.invoke_connector: Represents the request logic and execution on the
connector execution layer.These spans help you visualize end-to-end parent-child relationship workflows
from the assistant prompt to the third-party API. You can search and filter for
these spans in the Trace Explorer by service or span name, query them using
turn-level legacy assist tokens via the gemini_enterprise.assist_token
attribute, or look up traces using the W3C trace ID.
For more information, see Access traces and spans.
Process-level sandboxing and auto-updates (Preview)
This release introduces process-level sandboxing, automatic update checks, and other improvements to the CodeMender CLI (Preview):
config.yaml or per-command using the new --sandbox flag.cm update to forcefully check for and apply
updates immediately.cm stats subcommand to
view a summary of token usage (input, output, cached, thought, tool-use) for
your local sessions.For more information, see Install the CLI and configure.
Google Distributed Cloud (software only) for VMware 1.33.1100-gke.72 is now available for download. To upgrade, see Upgrade clusters. Google Distributed Cloud 1.33.1100-gke.72 runs on Kubernetes v1.33.11-gke.100.
If you use a third-party storage vendor, check the listing of our previously-qualified storage partners.
After a release, it takes approximately 7 to 14 days for the version to become available for use with GKE On-Prem API clients: the Google Cloud console, the gcloud CLI, and Terraform.
The following issues were fixed in 1.33.1100-gke.72:
gkectl prepare failed with a
permission denied error when attempting to read a private
registry CA certificate. The certificate file permissions
are now set to 644 so non-root processes can read it.
spec.selector fields on
existing ANG resources. The upgrade operator now preserves existing label
selectors during reconciliation so that V1 to V2 cluster migrations complete
successfully.
Google Distributed Cloud (software only) for bare metal 1.33.1100-gke.72 is now available for download. To upgrade, see Upgrade clusters. Google Distributed Cloud for bare metal 1.33.1100-gke.72 runs on Kubernetes v1.33.11-gke.100.
After a release, it takes approximately 7 to 14 days for the version to become available for installations or upgrades with the GKE On-Prem API clients: the Google Cloud console, the gcloud CLI, and Terraform.
If you use a third-party storage vendor, check the listing of our previously-qualified storage partners.
The following changes were added in 1.33.1100-gke.72:
csi-snapshot-validation-webhook component.
Upstream Kubernetes validation is now handled natively via Common
Expression Language (CEL) rules within the deployed Custom Resource
Definitions (CRDs). For more information, see Volume snapshots.
The following issues were fixed in 1.33.1100-gke.72:
Now available in preview, enhanced observability metrics — including engagement and estimated token usage data — are available for Conversational Analytics on the Conversational Analytics System Activity dashboard. To enable this feature, a Looker admin must turn on the Conversational Analytics Agent Token usage setting on the General page in the Preview section of the Admin panel.
Now available in preview, Looker admins can review end-user query success rates, rating distributions, and written user feedback in the Responses & Feedback tab on the Conversational Analytics System Activity dashboard. This data helps administrators improve system performance, support troubleshooting, and refine data agents.
To enable this feature, a Looker admin must turn on the End User Conversational Analytics (CA) Query Review setting on the General page in the Preview section of the Admin panel. The End User Conversational Analytics (CA) Query Review admin setting is disabled by default.
Specific user query data is visible only for users who have consented to share their query data in their account settings. The End User Conversational Analytics (CA) Query Review user setting is disabled by default.
From August 3 through August 7, 2026, the following features will be automatically enabled for Looker (original) and Looker (Google Cloud core) instances running Looker 26.12.
Preview: Semantic tags are available in
Preview.
Semantic tags provide standardized key-value metadata backed by OpenTelemetry
(OTel) conventions. Tags automatically replicates Environment and
Criticality attributes set on App Hub services and workloads as
read-only system semantic tags (google:AppHub/environment and
google:AppHub/criticality) on underlying direct resources. You can also view
available semantics and their OTel mappings in the Semantic Catalog in the
Google Cloud console.
For more information, see Tags overview and Create and manage tags.
Security Command Center released new Malicious Skill runtime threat detectors for Google Kubernetes Engine (GKE), Cloud Run, and Agent Platform. These detectors identify when a malicious skill (an AI agent capability) is executed or loaded. A malicious skill is any malicious binary that has been tagged as an LLM skill by Google's threat intelligence.
For more information, see the following:
General availability support for the following integration:
]]>Route LLM requests with model routing
You can now use model routing in API Gateway as a managed traffic management layer to accept OpenAI-compatible prompt requests, transcode them in-flight, and route them to specific foundation models in Gemini Enterprise Agent Platform Model Garden (including Gemini, Anthropic Claude, and OpenAI GPT models).
Key benefits and capabilities include:
x-google-api-management.ai.models.routing and x-google-model-router OpenAPI 3.x extensions.For more information, see Overview of model routing and Configure model routing.
The Assured Open Source Software Premium tier supports JavaScript (NPM) packages for Node.js environments.
For more information, see Download NPM packages using direct repository access and List of supported NPM packages for the Assured Open Source Software premium tier.
You can now change the backup plan associated with a Cloud SQL instance. This allows you to switch an instance to a different backup plan, provided the new plan uses the same backup vault and is in the same region as the instance. This feature is available through the Google Cloud console and gcloud CLI. To learn more, see Change the associated backup plan for a Cloud SQL instance.
You can now use cross-cloud connections to query data in AWS, Azure, and Salesforce Data 360 from all BigQuery regions. These connections let you use more BigQuery features and are more cost efficient than standard connections that use BigQuery Omni. This feature is in Preview.
The JDBC driver for BigQuery now supports OpenTelemetry for tracing and logging, which helps you monitor the performance of your database interactions and troubleshoot issues. Automatic exports to Google Cloud Observability are also available. This feature is generally available (GA).
Support for hybrid search (using the VECTOR_SEARCH function to combine a
semantic search with a lexical (keyword) search) has been restored. Using
HYBRID mode in the AI.SEARCH function has also been restored.
You can change the backup plan for your Cloud SQL enhanced backups without first removing the existing plan. For more information, see Change your instance's associated backup plan.
You can change the backup plan for your Cloud SQL enhanced backups without first removing the existing plan. For more information, see Change your instance's associated backup plan.
You can change the backup plan for your Cloud SQL enhanced backups without first removing the existing plan. For more information, see Change your instance's associated backup plan.
Updated the following JetBrains preconfigured base images to version 2026.x:
Changed: The minimum provisioned throughput for a Hyperdisk ML volume attached to more than 20 instances is 20 MiB/s per instance, reduced from 100 MiB/s.
For more information, see Share a Hyperdisk ML volume between instances.
Config Controller now uses the following versions of its included products:
Gemini Enterprise: New data stores and support for new actions (Public Preview)
The following data stores are in Public Preview:
Additionally, support for new actions is available in Public Preview for the following data stores:
TPU Subslicing (also known as Dynamic Subslicing) is now generally available for Ironwood (TPU7x). This feature enables you to incrementally provision node pools for a cube or litepod, breaking them into smaller slices (subslices) to run workloads requiring smaller topologies. Updates in this GA release include:
4x4x4, such as 2x2x1,
2x2x2, 2x2x4, and 2x4x4): Supported in GKE version
1.36.0-gke.3712000 or later.4x4x4 or larger): Supported in GKE
version 1.35.2-gke.1842000 or later.cloud.google.com/gke-tpu-partition-[shape]-state to specify smaller
subslice shapes. It also introduces UNSET and INCOMPLETE states. Support
for the DEGRADED state only applies to the top-level 4x4x4 topology, and
not for smaller sub-slicing topologies.For more information, see About GKE dynamic slicing.
[Spotlight Feature] Threat Hunt Agent
The Threat Hunt Agent is now available in Public Preview for Google SecOps Enterprise Plus customers. Powered by Gemini and grounded in Google Threat Intelligence (GTI), Mandiant frontline expertise, and the MITRE ATT&CK® framework, the Threat Hunt Agent autonomously automates proactive threat hunting across your historical security telemetry. For more information, see Threat Hunt Agent.
Key capabilities include:
The deadline for Stage 2 of the SOAR migration to Google Cloud has been extended from September 30th to November 30th, 2026. For more information, refer to the SOAR migration guide.
The deadline for Stage 2 of the SOAR migration to Google Cloud has been extended from September 30th to November 30th, 2026. For more information, refer to the SOAR migration guide.
Version 20260717.00 of the guest agent
is now available for Windows only. This version introduces the following fix:
Organization Policy Service custom constraints are available for Privileged Access Manager (PAM). You can use custom constraints to restrict how users create and modify entitlements and grants. This feature is in Preview.
For more information, see Use custom organization policies for Privileged Access Manager.
A localized processing option for Conversational Analytics is now available for European Union (EU) based customers, enabling data processing within the EU jurisdiction by routing traffic through an EU multi-regional endpoint.
Note that the EU multi-region endpoint covers data residency strictly within
EU member states and does not include the europe-west2 (London, UK) or
europe-west6 (Zurich, Switzerland) regions. Support for the europe-west12 (Turin, Italy)
region is also unavailable at this time.
Connectivity Tests supports testing connectivity from a Database Migration Service private connection to a Cloud SQL instance.
For more information, see Test from a Database Migration Service private connection to a Cloud SQL instance.
The Assured Open Source Software Premium tier supports JavaScript (NPM) packages for Node.js environments.
For more information, see Download NPM packages using direct repository access and List of supported NPM packages for the Assured Open Source Software premium tier.
For the Security Command Center Standard tier, AI Protection is supported for both projects and organizations.
Project-level activations for the Standard tier include access to the AI security dashboard, basic inventory view (excluding Gemini models), and baseline security findings.
Some features of AI Protection are only available for the Premium and Enterprise tiers or for organization-level activations. For more information, see Configure AI Protection.
Preview stage support for the following integration:
Preview: You can create v2 IPv4 public advertised prefixes for bring your own IP addresses (BYOIP) that use Standard Tier IP addresses. For more information, see Network Service Tiers.
]]>Fixed issue with JupyterLab UI silently reverting file changes and interrupting kernels due to Gemini CLI's auto reload extension.
The JupyterLab last-active and auto-reload extensions are now enabled only when Gemini CLI is configured, preventing them from affecting the native JupyterLab UI otherwise.
Installed latest packages from upstream dependencies.
Fixed issue with JupyterLab UI silently reverting file changes and interrupting kernels due to Gemini CLI's auto reload extension.
The JupyterLab last-active and auto-reload extensions are now enabled only when Gemini CLI is configured, preventing them from affecting the native JupyterLab UI otherwise.
Installed latest packages from upstream dependencies.
Fixed an error that could prevent creating new micromamba environments at runtime due to package cache permissions.
Fixed an error that could prevent creating new micromamba environments at runtime due to package cache permissions.
Fixed issue with JupyterLab UI silently reverting file changes and interrupting kernels due to Gemini CLI's auto reload extension.
The JupyterLab last-active and auto-reload extensions are now enabled only when Gemini CLI is configured, preventing them from affecting the native JupyterLab UI otherwise.
Installed latest packages from upstream dependencies.
Fixed issue with JupyterLab UI silently reverting file changes and interrupting kernels due to Gemini CLI's auto reload extension.
The JupyterLab last-active and auto-reload extensions are now enabled only when Gemini CLI is configured, preventing them from affecting the native JupyterLab UI otherwise.
Installed latest packages from upstream dependencies.
Release 6.3.96 is being rolled out to the first phase of regions as listed here.
This release contains internal and customer bug fixes.
]]>Gemini Enterprise: Pay for what you consume with the Pay-as-you-go edition
The Gemini Enterprise Pay-as-you-go edition is generally available. With this edition, you pay for the feature usage your project consumes without pooled user license quotas. To get started, you need an invoiced Cloud Billing account with a one-seat minimum. You can monitor feature usage and set monthly spend limits in the Google Cloud console. This feature is rolling out gradually to eligible customers over the next few weeks.
For more information, see the following:
Preview: Google Cloud VMware Engine capacity allocations are available in preview. A capacity allocation is a global reservation of physical nodes for your Cloud Billing account. Your reserved nodes are organized into placement groups (PGs), which represent the specific physical hardware and location of your reserved capacity.
Google Cloud VMware Engine capacity allocations let you do the following: * Use a capacity allocation during private cloud creation by selecting a PG. * Use a capacity allocation when adding a cluster. * Configure a hybrid private cloud across multiple PGs.
For more information, see Manage capacity allocations.
Release 6.3.95 is now available for all regions.
]]>App Topology (including One Graph) is generally available (GA).
App Topology (including One Graph) is generally available (GA).
On July 31, 2026 we released an updated version of the Apigee hybrid software, v1.15.7.
Runtime rollout strategy configuration
In this release, you can configure the rollout strategy used when updating runtime (message processor) ReplicaSets by setting the runtime.release.strategy property (with options rolling, scale-down-first, or none) or per-environment with envs[].components.runtime.release.strategy in your overrides configuration file. The property defaults to rolling.
Various security and CVE fixes are included in this release.
Cloud Load Balancing introduces a new version of the Network Load Balancer—the global external passthrough Network Load Balancer, which is the global variant of the regional external passthrough Network Load Balancer. The load balancer is available in Preview.
This load balancer variant solves use cases for Security Service Edge (SSE), DNS hosting, Adtech (real-time bidding), real-time communications (RTC), live streaming, and online gaming, among others.
Global external passthrough Network Load Balancers are Layer 4 passthrough load balancers that distribute external traffic among backends (instance groups or network endpoint groups) that can reside in multiple Google Cloud regions. By using Google's global anycast IP routing, the global external passthrough Network Load Balancer steers user traffic to the closest region with healthy backends and available capacity, delivering ultra-low latency and dynamic cross-region failover to ensure resilience to regional outages.
The load balancer provides you with two external IP addresses, each served by a disjoint and isolated global load balancing control and data plane server infrastructure (also known as an availability group) to provide high availability.
The load balancer supports TCP, UDP, ESP, GRE, ICMP, and ICMPv6 traffic and can handle both IPv4 and IPv6 traffic. You can deploy your backends in any of the following Google Cloud regions:
us-west1, us-west4, us-east4, us-east5europe-west2, europe-west3asia-southeast1, asia-south1, asia-northeast1southamerica-east1africa-south1australia-southeast1Note that this release doesn't support GKE backends for the global external passthrough Network Load Balancer.
For details on the new load balancer, see Global external passthrough Network Load Balancer overview.
Starting on August 1, 2026, when you create or clone a Cloud SQL instance enabled with Private Service Connect, or when you enable Private Service Connect for an existing instance, then connection reconciliation behavior is enabled by default and can't be disabled.
When you remove a project from the list of allowed projects, all existing Private Service Connect connections from the removed project are immediately closed (reconciled). This means that applications using Private Service Connect endpoints in those removed projects can't continue to connect to the Cloud SQL instance using those endpoints.
For more information, see Allowed Private Service Connect projects.
QueryData adds support for parameterized secure views (PSVs) to help secure applications that use natural language queries. For more information, see Secure and control access to application data.
This feature is in Preview.
Starting on August 1, 2026, when you create or clone a Cloud SQL instance enabled with Private Service Connect, or when you enable Private Service Connect for an existing instance, then connection reconciliation behavior is enabled by default and can't be disabled.
When you remove a project from the list of allowed projects, all existing Private Service Connect connections from the removed project are immediately closed (reconciled). This means that applications using Private Service Connect endpoints in those removed projects can't continue to connect to the Cloud SQL instance using those endpoints.
For more information, see Allowed Private Service Connect projects.
Starting on August 1, 2026, when you create or clone a Cloud SQL instance enabled with Private Service Connect, or when you enable Private Service Connect for an existing instance, then connection reconciliation behavior is enabled by default and can't be disabled.
When you remove a project from the list of allowed projects, all existing Private Service Connect connections from the removed project are immediately closed (reconciled). This means that applications using Private Service Connect endpoints in those removed projects can't continue to connect to the Cloud SQL instance using those endpoints.
For more information, see Allowed Private Service Connect projects.
Cloud SQL for SQL Server now supports executing SQL statements using the Cloud SQL Data API.
Support for Intel TDX on c4-standard-* machine types is
available in Preview.
You can now pause a Dataflow batch job using the pause_on_failure
service option. This feature lets you preserve the state of your batch pipeline
job, address external issues, and resume processing without losing completed
work. You can use this option to automatically pause a job on failure, or use
this feature to manually pause a job when you chose to. For more information,
see Pause a Dataflow job.
You can now replicate change data from Workday with Datastream. For more information, see Stream data from Workday.
This feature is in Preview.
You can generate synthetic data for a Managed Service for Apache Kafka cluster by using Dataflow. For more information, see Generate synthetic data for a Managed Service for Apache Kafka cluster.
You can use custom organization policies to improve the security, compliance, and governance of your Memorystore for Valkey instances by enforcing consistent configurations and restrictions for the instances. This ensures that your instances adhere to security best practices and regulatory requirements. This feature is Generally Available.
Organizations that are enrolled in the data residency Preview program can update their organization's data residency and data encryption configuration. For more information, see Modify data residency or data encryption configuration.
Agent Platform Vulnerability Assessment (Preview) scans for plaintext secrets, such as credentials, access tokens, and API keys, in customer-deployed Gemini Enterprise Agent Platform containers. For more information, see Agent Platform Vulnerability Assessment.
]]>Installed latest packages from upstream dependencies.
Installed latest packages from upstream dependencies.
Fixed issue with JupyterLab UI silently reverting file changes and interrupting kernels due to Gemini CLI's auto reload extension.
The JupyterLab last-active and auto-reload extensions are now enabled only when Gemini CLI is configured, preventing them from affecting the native JupyterLab UI otherwise.
Fixed the JupyterLab Git extension's Pull and Push buttons being disabled by pinning jupyterlab-git to 0.53.0.
Fixed the JupyterLab Git extension's Pull and Push buttons being disabled by pinning jupyterlab-git to 0.53.0.
Fixed an error that could prevent creating new micromamba environments at runtime due to package cache permissions.
Fixed an error that could prevent creating new micromamba environments at runtime due to package cache permissions.
Workbench internal agents now trust custom and enterprise CA certificates installed on the host operating system, fixing TLS certificate verification failures for connections routed through the instance proxy.
Workbench internal agents now trust custom and enterprise CA certificates installed on the host operating system, fixing TLS certificate verification failures for connections routed through the instance proxy.
Files created by the post-startup script are now owned by the jupyter user, and git safe.directory is configured so that root-owned repositories continue to work.
Files created by the post-startup script are now owned by the jupyter user, and git safe.directory is configured so that root-owned repositories continue to work.
Fixed issue with JupyterLab UI silently reverting file changes and interrupting kernels due to Gemini CLI's auto reload extension.
The JupyterLab last-active and auto-reload extensions are now enabled only when Gemini CLI is configured, preventing them from affecting the native JupyterLab UI otherwise.
The JupyterLab last-active and auto-reload extensions are now enabled only when Gemini CLI is configured, preventing them from affecting the native JupyterLab UI otherwise.
You can add tables, views, data sources, and data quality tests as tasks to BigQuery pipelines. For more information, see Add a pipeline task. This feature is generally available.
You can audit when users download query results by using the BigQuery
console. Data Access audit logs for the tabledata.list method now include a
uiDownloadRequest
field to indicate whether the request was triggered by a UI download.
This feature is
generally available
(GA).
You can use Bigtable as a remote storage backend for LMCache. By storing the large language model (LLM) key-value (KV) cache externally in Bigtable, multiple AI serving instances can share and reuse precomputed attention tensors. This reduces compute overhead and significantly improves time-to-first-token (TTFT) for repeated prompts and shared documents. This feature is in Preview.
Support for retrieving, listing, enabling, and disabling Model Context Protocol (MCP) servers and tools using the Cloud API Registry API is shut down. For supported services, MCP endpoints are available once you enable the API. For more information, see Feature deprecations.
Config Connector version 1.154.1 is now available.
New Alpha Resources (Direct Reconciler):
ApigeeApiProduct
ApigeeRegistryApi
ApigeeRegistryArtifact
APIHubExternalAPI
APIHubInstance
AppOptimizeReport
ArtifactRegistryVPCSCConfig
BigQueryMigrationMigrationWorkflow
BlockchainNodeEngineBlockchainNode
CCInsightsConversation
CCInsightsIssueModel
CCInsightsPhraseMatcher
CESApp
CloudBuildConnection
CloudSecurityComplianceFramework
ConnectorsConnection
ContentWarehouseDocument
ContentWarehouseRuleSet
ContentWarehouseSynonymSet
DatabaseMigrationPrivateConnection
DataformFolder
DataformTeamFolder
DataLabelingDataset
DataLabelingEvaluationJob
DataLineageProcess
DataplexAspectType
DataplexDataAttributeBinding
DataplexDataScan
DataplexDataTaxonomy
DataplexGlossary
DataplexMetadataJob
DevConnectConnection
DialogflowConversationDataset
DialogflowSecuritySettings
DialogflowSipTrunk
DiscoveryEngineControl
DiscoveryEngineSampleQuerySet
DLPConnection
DLPDiscoveryConfig
EventarcGoogleApiSource
GeminiDataAnalyticsConversation
GKEBackupBackupChannel
LiveStreamAsset
ManagedKafkaConnectCluster
MigrationCenterGroup
NetworkSecurityAddressGroup
NetworkSecurityAuthzPolicy
NetworkSecurityFirewallEndpoint
NetworkSecurityFirewallEndpointAssociation
NetworkSecurityGatewaySecurityPolicy
NetworkSecurityPartnerSSEGateway
NetworkSecurityPartnerSSERealm
NetworkSecuritySecurityProfile
NetworkSecurityTLSInspectionPolicy
NetworkServicesAuthzExtension
NotebooksSchedule
RedisClusterEndpoint
RunWorkerPool
SaasServiceMgmtRelease
SQLAdminBackup
StorageInsightsDatasetConfig
TestingDeviceSession
TranslateAdaptiveMtDataset
VectorSearchCollection
VertexAIFeatureGroup
VertexAIFeatureOnlineStore
VertexAIPipelineJob
VertexAISpecialistPool
VertexAIStudy
VertexAITuningJob
VideoStitcherCDNKey
VisionProduct
VMwareEnginePrivateConnection
New Fields:
ComputeSubnetwork
spec.reservedInternalRange field.NetworkConnectivityInternalRange
spec.allocationOptions field.ComputeNetwork
spec.networkProfile field.ComputeSecurityPolicy
spec.region field.DNSRecordSet
healthCheckRef and rrdatasRefs fields.ComputeAddress
spec.ipCollection field.ComputeURLMap
spec.defaultCustomErrorResponsePolicy field.spec.test[].expectedOutputUrl and spec.test[].expectedRedirectResponseCode fields.RedisCluster
spec.crossClusterReplicationConfig field.MonitoringAlertPolicy
spec.conditions[].conditionSql field (SQL Condition).StorageBucket
spec.ipFilter field.PubSubTopic
spec.messageStoragePolicy.enforceInTransit field.ComputeRouterNAT
Reconciliation Improvements:
We have added support for direct reconciliation to more resources, with opt-in behaviour. The API is unchanged. To use the direct reconciler, add the cnrm.cloud.google.com/reconciler: direct annotation to the corresponding Config Connector object.
BigQueryReservationCapacityCommitmentBigtableGCPolicyBillingBudgetsBudgetCertificateManagerCertificateMapCertificateManagerCertificateMapEntryComputeAddressComputeAutoscalerComputeBackendServiceSignedURLKeyComputeDiskComputeDiskResourcePolicyAttachmentComputeExternalVPNGatewayComputeFirewallComputeFirewallPolicyComputeHTTPHealthCheckComputeHTTPSHealthCheckComputeImageComputeInstanceComputeInstanceGroupComputeInstanceGroupManagerComputeNetworkComputeNodeTemplateComputeRouteComputeRouterComputeRouterInterfaceComputeRouterNATComputeSSLPolicyComputeSecurityPolicyComputeTargetHTTPSProxyComputeURLMapDataflowJobDataprocAutoscalingPolicyDataprocClusterDNSResponsePolicyKMSCryptoKeyKMSKeyRingLoggingLogExclusionMonitoringAlertPolicyNetworkServicesGatewayPrivateCACertificateAuthorityPrivateCACertificateTemplatePubSubSubscriptionRecaptchaEnterpriseKeyRedisInstanceServiceDirectoryEndpointServiceDirectoryNamespaceServiceServiceIdentityBug Fixes:
ComposerEnvironment: Fix storageConfig.bucketRef mapping.MemorystoreInstance: Prevent infinite reconciliation drift loop by aligning connections list length.MemorystoreInstance: Prevent false drift and update attempts on unspecified immutable fields.ComputeBackendService: Fix config-connector export tool to export backend field.KMSAutokeyConfig: Clean up and improve autokey config identity and deletion resolution.BigQuery: Fix perpetual diff on tables inheriting dataset encryption.NotebooksInstance: Fix direct controller for NotebookInstance to resolve references.| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.6.143 | v27.5.1 | v2.0.10 | See List |
Added opt-in support for the guest agent extensions manager, MWLID, and Telemetry extensions.
Fixed CVE-2026-29111 in sys-apps/systemd
Fixed CVE-2026-35177 in app-editors/vim and app-editors/vim-core.
Fixed CVE-2026-3644 in dev-lang/python
Fixed CVE-2026-40355 and CVE-2026-40356 in app-crypt/mit-krb5.
Fixed CVE-2026-53381 in the Linux kernel.
Fixed CVE-2026-53385 in the Linux kernel.
Fixed CVE-2026-53388 in the Linux kernel.
Fixed CVE-2026-53391 in the Linux kernel.
Fixed CVE-2026-53397 in the Linux kernel.
Fixed CVE-2026-53398 in the Linux kernel.
Fixed CVE-2026-59995, CVE-2026-59996, CVE-2026-59997, CVE-2026-59999, CVE-2026-60000, CVE-2026-60001, and CVE-2026-60002 in openssh.
Fixed CVE-2026-6019 in dev-lang/python
Fixed CVE-2026-63794 in the Linux kernel.
Fixed CVE-2026-63795 in the Linux kernel.
Fixed CVE-2026-63800 in the Linux kernel.
Fixed CVE-2026-63802 in the Linux kernel.
Fixed CVE-2026-63807 in the Linux kernel.
Fixed CVE-2026-63809 in the Linux kernel.
Fixed CVE-2026-63823 in the Linux kernel.
Fixed CVE-2026-63824 in the Linux kernel.
Fixed CVE-2026-63827 in the Linux kernel.
Fixed CVE-2026-63828 in the Linux kernel.
Fixed CVE-2026-63830 in the Linux kernel.
Fixed CVE-2026-64244 in the Linux kernel.
Fixed CVE-2026-64247 in the Linux kernel.
Upgraded net-libs/nghttp2 to 1.69.0 and fixed CVE-2026-58055.
Google Cloud Cortex Framework version 7 is now generally available. Version 7 introduces a modular deployment architecture, simplified data orchestration via Dataform, and AI-ready data products with BigQuery and Knowledge Catalog integration. This enables enterprises to build, extend, and deploy data assets and pipelines for advanced analytics and agentic use cases with less risk, complexity, and cost.
New features and enhancements
additional to those released in preview
AI, discovery, and governance:
Expanded data product content and integrations:
Supportability and operations:
Google Cloud Cortex Framework version 7 includes telemetry to capture anonymized deployment statistics. This data helps the solution build team focus on improving modules with high adoption. Telemetry is enabled by default, but you can opt out at any time.
The following features are rolling out over the next week.
Conversational Analytics is generally available
Conversational Analytics in Data Studio is now generally available. You can now filter your data agents by the Google Cloud project to which they belong. Agents that require additional permissions are now displayed with an Unavailable label.
Email notifications when sharing Conversational Analytics data agents
When you share data agents that were created in BigQuery with Data Studio users, you can opt to send an email to notify those users of their access to the agent.
Google Cloud CCaaS 5.2
We've released version 5.2 of Google Cloud CCaaS.
The timing of the update to your instance depends on the deployment schedule that you have chosen. For more information, see Deployment schedules.
Answering machine detection for progressive campaigns
Answering machine detection (AMD) is now supported for progressive outbound campaigns. When enabled, Contact Center AI Platform analyzes call audio in the background to determine whether a call reaches a live person, or whether it reaches an answering machine or voicemail. If an answering machine or voicemail is detected, the call ends and the dialer proceeds to the next contact.
Administrators: In the Settings > Campaigns > Dialer Modes pane, there's a new Enable Answering Machine Detection toggle.
User experience change: When CCAI Platform detects voicemail or an answering machine, a green banner appears in the call adapter to indicate this.
For more information, see Answering machine detection for progressive campaigns.
New endpoints for getting email sessions and messages
Two new read-only endpoints are now available, allowing external systems to retrieve the parsed contents of an email interaction. This includes the sender, recipients, subject, body, and attachment metadata. Here are the endpoints:
/apps/api/v1/email/sessions/EMAIL_SUPPORT_ID: Returns
email session summary information and a list of message IDs with metadata.
/apps/api/v1/email/messages/EMAIL_THREAD_ID: Returns the
full content of a single message.
For more information, see Get email sessions and messages.
Email forwarding with attachments
Agents can now forward emails to external recipients directly from the email adapter. When forwarding, all attachments from the original email are automatically included. Agents can remove attachments before sending, if needed. The original email remains in its assigned queue with its status unchanged.
User experience change: A new Forward button is available in the email adapter.
For more information, see Forward an email.
Smart disposition
Smart disposition is a new AI-powered capability that automatically suggests a disposition code at the end of a session. This reduces manual work for agents and improves data consistency.
Administrators: There's a new Smart Disposition toggle in the following locations:
The Settings > Operation Management > Wrap-up > Automatic wrap-up for inbound calls > Disposition Codes & Notes for calls > Disposition Codes section.
The Settings > Operation Management > Wrap-up > Automatic wrap-up for outbound calls > Disposition Codes & Notes for calls > Disposition Codes section.
The Settings > Operation Management > Wrap-up > Automatic wrap-up for chats > Disposition Codes & Notes for chats > Disposition Codes section.
User experience change: When smart disposition is turned on, a suggested disposition displays in the Disposition field of the Wrap-up screen in the agent adapter.
For more information, see Smart disposition.
This release addresses the following issues:
Fixed an issue where voice calls became stuck in a virtual agent state after a session ended abnormally or an escalation handoff didn't complete.
Fixed an issue where using invalid sorting parameters on agent activity logs caused a server error.
Fixed an issue where canceling a warm transfer to a queue at overcapacity caused the caller to become stranded in an automated menu loop.
Fixed an issue where the agent activity logs API allowed unbounded time ranges, which led to system performance degradation and gateway timeouts.
Fixed an issue where the team and menu endpoints experienced performance delays.
Fixed an issue where calls transferred to a queue with overcapacity deflection (OCD) enabled didn't redirect, causing callers to hear indefinite ringing.
Fixed an issue where the call duration in the call adapter and in the in-call status timer of the agent desktop didn't match for outbound calls.
Fixed an issue where the Agent Activity dashboard displayed time zones inconsistently.
Fixed an issue where the inactivity timer didn't force a session to end if the agent closed the browser or browser tab before the inactivity timer expired.
Fixed an issue where the All Call History and Individual Call History reports displayed incorrect cascade group numbers for transferred calls and chats.
Fixed an issue where the sentiment score appeared in the Call Details panel of the agent desktop despite sentiment analysis being turned off in the conversation profile.
Fixed an issue with IVR calls routed through Nexmo, where the virtual agent missed the first several seconds of caller audio, forcing callers to repeat themselves.
Fixed an issue where chats became stuck in the queue after a deltacast routing projection expired or an agent didn't connect.
Fixed an issue where reordering queues in the CCAI Platform portal caused significant latency and required a manual page refresh to display the changes.
Fixed an issue with Salesforce integrations where adding a third party to a call incorrectly displayed their contact name in the customer field in the agent desktop Participants panel.
Fixed an issue where direct inbound SMS chats that were sent to an unavailable agent expired and failed instead of being rerouted.
Fixed an issue where anonymous inbound calls incorrectly displayed an agent's contact information instead of indicating an unknown caller.
Fixed an issue where database deadlocks caused transactions to run against the wrong database, causing failed lookups and chat sessions getting stuck.
Fixed an issue where manual status changes to Available after an automatic wrap-up were incorrectly attributed to the system instead of the agent in activity reports.
Fixed an issue where retrieving large datasets using the manager API caused connection timeouts and incomplete data synchronization for downstream systems.
Fixed an issue where the Agent Assist Hub incorrectly displayed The
Agent Assist Hub feature is not enabled during voice sessions despite the
Agent Assist Hub being enabled.
Fixed a web SDK issue where navigating between pages on the host website during an active chat resulted in a duplicate session being created.
Fixed an issue where the AgentSystemData historical report incorrectly showed zero login time for agents who were actively handling calls.
Fixed an issue where Dialogflow agents incorrectly escalated or disconnected calls when responding with only pre-recorded audio.
Fixed an issue where transfer completion events appeared twice in the agent desktop session data feed when a user was transferred to a task virtual agent.
Fixed an issue where virtual agent interactions failed and escalated prematurely.
Fixed an issue where a single wrap-up event was incorrectly dispatched twice to Dialogflow on bidirectional-enabled voice calls.
Fixed an issue where session summarization sections weren't displayed in the order that they were configured.
Fixed an issue where sensitive information in the CC and BCC fields of email requests was visible in system logs.
Fixed an issue where the email adapter displayed a blank gray screen when an agent attempted to transfer emails between queues.
Fixed an issue in Kustomer integrations where abandoned or failed calls
weren't finalized, leaving records in the Call In Progress state.
Fixed an issue in ServiceNow integrations where starting a chat from a queue caused duplicate cases for a single chat ID.
Fixed an issue where mobile chat sessions ended unexpectedly after successfully escalating to a human agent.
Fixed an issue where chat transcript PDF headers remained in English for non-English queues.
Fixed an issue where placeholder text (Content cards displayed here)
didn't appear in the agent adapter or live chat view when content cards
weren't supported.
Fixed an issue where agents received duplicate SMS messages from end-users.
Fixed an issue where the inactivity timeout didn't trigger for chats waiting in a transfer queue.
Fixed an issue where chats waiting in a transfer queue remained open indefinitely and created duplicate metadata files.
Fixed an issue where chat transcripts were missing from data exports when a CRM ticket wasn't created.
Fixed an issue where chats escalated from a virtual agent remained in queued status after being assigned to an agent.
Fixed an issue where calls escalated from a virtual agent to a holiday-closed queue bypassed the holiday message and played the after-hours deflection message.
Fixed an issue where call recording URLs weren't saved or synced to the CRM for calls with multiple audio segments, such as those involving an IVR followed by an agent conference.
Fixed an issue where voice calls remained in an assigned state without progressing or requeueing if a system error occurred during the assignment process.
Fixed an issue where callers were disconnected from voicemail greetings after an agent completed their wrap-up following a call transfer.
Fixed an issue where manual wrap-up session data was incorrectly recorded across multiple sessions, leading to inflated duration reports.
Fixed an issue where duplicate call recording files with an .N suffix were
created in external storage.
Fixed an issue where calls prematurely disconnected or experienced audio loss during the hangup process.
Fixed an issue where Telnyx VoIP calls silently dropped without notification if the connection was interrupted.
Improved internal instrumentation to diagnose poor call quality scores and improve call reliability.
Fixed an issue where agents configured for Deltacast and auto-answer received unexpected multicast call offers.
Fixed an issue where the agent desktop didn't save call-related settings such as mute status.
Fixed an issue where queue-level automatic wrap-up settings were unexpectedly disabled.
Fixed an issue where the switch to chat button wasn't accessible from the keyboard.
Fixed an issue where an agent ended the wrap-up session of another agent.
Fixed an issue where agents experienced delays of up to 30 seconds when transitioning from "Wrap" to "Available" status.
Fixed an issue where an An error has occurred message incorrectly appeared
when a user created a direct access point for instances without a configured
CRM.
Fixed an issue where message timestamps weren't visible to supervisors and administrators when monitoring active chats in the CCAI Platform portal.
Fixed an issue where historical data syncs to Calabrio failed.
Fixed an issue where the bulk user upload process allowed unauthorized role assignments.
Fixed an issue where deleting a notification rule removed all associated historical notifications without warning.
Fixed an issue where call events in timeline reports appeared out of chronological order.
Fixed an issue where technical connection timeouts during call offers weren't correctly tracked as skipped interactions.
Fixed an issue where chats remained in the queue following a virtual agent escalation.
Fixed an issue where importing a contact list CSV for a native power dial campaign failed if the file contained malformed rows or encoding issues.
We've added the city field to Customer Insights reports and Detailed
Disbursements reports.
For more information, see Customer Insights report fields and Detailed Disbursements report fields.
GKE cluster versions have been updated.
New versions available for upgrades and new clusters.
The following versions are now available for new GKE clusters, and for manual control plane upgrades and node upgrades for existing clusters. For more information about versioning and upgrades, see GKE versioning and support and About GKE cluster upgrades.
Google Chronicle: Version 91.0
Updated Wiz Defend alert naming format in the following connector:
From July 27 through July 30, 2026, the following features will be automatically enabled for Looker (original) instances running Looker 26.12.
Looker admins now have the ability to configure a Looker instance to require multi-factor authentication (MFA) whenever a user tries to log in by using an email and a password. This feature is enabled by default.
The custom calendar feature is now generally available.
Localization is now supported for the dimension_group parameter. You can localize the timeframes, intervals, or custom timeframes generated by a dimension group by providing translations in your locale strings files. For more details, see Localizing dimension groups.
Note: This item was added on August 5, 2026.
The Expression Assistant is now generally available.
The Enhanced search feature is now generally available.
Looker Continuous Integration (CI) now supports email alerts. When you create or edit a CI suite, you can enable the Enable email alerts toggle to specify email recipients and select which run statuses will trigger emails (Failed, Error, Passed, or Cancelled). For more information, see Set up alerting.
The LookML Projects page has been updated with a more performant tabbed layout, which features three tabs: Models and Projects, Pending Projects, and Marketplace Projects.
Now available in preview, the new Modern User Interface feature enables modernized layouts and design alongside new configuration settings for visualizations and dashboards. When this preview feature is enabled, users can apply a Modern visualization theme that features updated typography and modern, accessible color palettes for improved data legibility. Additionally, a new Modern dashboard style provides a high-density, streamlined design that optimizes data viewing and aligns with Google's latest design standards.
Now available in preview, verified queries (also referred to as golden queries) are predefined pairs of natural language questions and their exact, corresponding Looker Explore queries that act as verified standards of truth to teach your Explore data agent how to handle complex business requests without guessing.
To enable verified queries, a Looker admin must turn on the Verified Queries setting on the Gemini in Looker admin page. The Conversational Analytics setting must also be enabled for verified queries to be used.
When you chat in Gemini Enterprise with data agents that you create in Looker, agent responses now include charts and visualizations.
The Insight Assistant now displays the process the assistant uses to generate the response, showing key details in your data that it used to generate the response, and listing the fields from your Explore that it used.
Key updates in this image version include:
3.5-dataproc-28 subminor image version doesn't have preconfigured Conda channels, and is mapped to default aliases (such as 3.5 and latest).
3.5-dataproc-28 or using default aliases (3.5, latest), packages cannot be installed using Conda unless channels are manually configured during cluster initialization.You may need to delete and replace existing clusters After August 25, 2026, existing clusters created with images that have preconfigured Conda channels (even if cluster jobs don't use Conda to install packages) need to be deleted and replaced with new clusters created or recreated with images that don't have preconfigured Conda channels.
Oracle Database@Google Cloud supports customer-managed encryption keys (CMEK) for Exascale VM Clusters. You can enable CMEK on Exascale VM Clusters. This feature is generally available (GA).
Support for managing Model Context Protocol (MCP) endpoint enablement using the Service Usage v2beta API is shut down.
For more information, see Feature deprecations.
General Availability: You can use the Resolve subnet mask setting on a
subnet to configure all attached Compute Engine instances with the same netmask
as the subnet (instead of /32). Configuring larger instance netmasks lets
compute instances discover the MAC addresses of other machines within the same
subnet and directly communicate with them by using destination MAC addresses.
For more information, see Compute instance netmasks.
]]>Support for enabling only needed legacy bundled services using the
app_engine_bundled_services
field is in General Availability.
Support for enabling only needed legacy bundled services using the
app_engine_bundled_services
field is in General Availability.
Support for enabling only needed legacy bundled services using the
app_engine_bundled_services
field is in General Availability.
Support for enabling only needed legacy bundled services using the
app_engine_bundled_services
field is in General Availability.
The BigQuery Data Transfer Service now supports incremental data transfers when transferring data from Klaviyo to BigQuery. This feature is supported in Preview.
You can use the Google Cloud console to manage row key schemas for your Bigtable tables. This feature is in Preview.
Cloud KMS Autokey with same-project key storage (formerly known as Autokey for delegated key management) is generally available. Autokey with same-project key storage can be used on its own or alongside Autokey with dedicated-project key storage (formerly known as Autokey for centralized key management).
For more information, see Enable Cloud KMS Autokey. To learn how to set guardrails to constrain how Autokey is used in your organization, see Control Autokey usage.
Cloud SQL for MySQL now supports significantly faster re-encryption of instances and replicas protected by customer-managed encryption keys (CMEKs), and re-encryption now completes with zero downtime. The steps to re-encrypt your instances and replicas are unchanged, but the operation now re-encrypts the underlying disks in-place, without creating re-encryption backups.
For more information, see Re-encrypt an existing CMEK-enabled instance or replica.
Cloud SQL for PostgreSQL now supports significantly faster re-encryption of instances and replicas protected by customer-managed encryption keys (CMEKs), and re-encryption now completes with zero downtime. The steps to re-encrypt your instances and replicas are unchanged, but the operation now re-encrypts the underlying disks in-place, without creating re-encryption backups.
For more information, see Re-encrypt an existing CMEK-enabled instance or replica.
Cloud SQL for SQL Server now supports significantly faster re-encryption of instances and replicas protected by customer-managed encryption keys (CMEKs), and re-encryption now completes with zero downtime. The steps to re-encrypt your instances and replicas are unchanged, but the operation now re-encrypts the underlying disks in-place, without creating re-encryption backups.
For more information, see Re-encrypt an existing CMEK-enabled instance or replica.
Cloud Scheduler is available in the following locations:
europe-west8 (Milan, Italy)europe-west9 (Paris, France)us-south1 (Dallas, United States)For the clusters using TRAFFIC_DIRECTOR implementation,
IP auto-allocation
with DNS Proxy is now supported in Rapid release channel.
Public preview: you can create a machine image that includes or excludes specific non-boot disks attached to a source Compute Engine instance. For more information, see Create a machine image from specific disks. You can also create a new disk by restoring a specific individual disk from a machine image without creating a new Compute Engine instance. For more information, see Create a disk from a machine image.
Confidential VM instances with AMD SEV on C3D and C4D machine types now support configurations with more than 255 vCPUs.
You can now create a Datastream stream directly from the instance or database overview page in Spanner using the automated flow.
For more information, see Create a Spanner stream using the automated flow.
Gemini Enterprise: New data stores and support for new actions (Preview)
The following data stores are available in Public Preview:
Additionally, support for new actions is available in Public Preview for the following data stores:
For more information, see Connect a third-party data source.
Google SecOps has updated the list of supported default parsers. Parsers are updated gradually, so it might take one to four days before you see the changes reflected in your region.
The following supported default parsers have been updated. Each parser is listed by product name and log_type value, where applicable. This list includes both released default parsers and pending parser updates.
AIRLOCK_DIGITAL)AIX_SYSTEM)AKAMAI_DATASTREAM_2)AKAMAI_SIEM_CONNECTOR)APACHE)ARCSIGHT_CEF)ARMIS_ALERTS)ARUBA_SWITCH)ATLASSIAN_AUDIT)AUDITD)AVAYA_AURA)AWS_CLOUDTRAIL)AWS_CLOUDWATCH)AWS_CONTROL_TOWER)AZURE_ACTIVITY)AZURE_AD)AZURE_AD_CONTEXT)AZURE_GATEWAY)AZURE_KEYVAULT_AUDIT)AZURE_RESOURCE_LOGS)BLUECOAT_WEBPROXY)BOMGAR)CATO_NETWORKS)CHECKPOINT_FIREWALL)CHECKPOINT_HARMONY)CHROME_MANAGEMENT)CHROMEOS_XDR)CISCO_ASA_FIREWALL)CISCO_EMAIL_SECURITY)CISCO_FIREPOWER_FIREWALL)CISCO_FIRESIGHT)CISCO_ISE)CISCO_ROUTER)CISCO_SWITCH)CISCO_UCM)CLAROTY_XDOME)CLAUDE_COMPLIANCE_LOGS)CLEARPASS)CLOUDFLARE)CORTEX_XDR)CS_EDR)DARKTRACE)EFFICIENTIP_DDI)F5_ASM)F5_BIGIP_LTM)FASTLY_CDN)FIREEYE_EMPS)FIREEYE_HX)FIREEYE_NX)FORCEPOINT_WEBPROXY)FORTINET_FIREWALL)FORTINET_FORTIANALYZER)FORTINET_FORTICLIENT)FORTINET_SWITCH)GCP_CLOUDAUDIT)GCP_SECURITYCENTER_EXTERNAL_EXPOSURE)GITLAB)GTI_IOC)GUARDDUTY)HUAWEI_SWITCH)IBM_SAM)IIS)ILLUMIO_CORE)IMPERVA_SECURESPHERE)INFOBLOX)INFOBLOX_DHCP)JAMF_PRO_CONTEXT)LOOKOUT_MOBILE_ENDPOINT_SECURITY)MACOS)MCAFEE_IPS)MICROFOCUS_IMANAGER)MICROSOFT_DEFENDER_ENDPOINT)MICROSOFT_DEFENDER_MAIL)MICROSOFT_GRAPH_ALERT)MICROSOFT_SENTINEL)MICROSOFT_SQL)MIMECAST_URL_LOGS)MISP_IOC)NETAPP_ONTAP)NETSKOPE_ALERT_V2)NIX_SYSTEM)OFFICE_365)OKTA)ONAPSIS)OPEN_VPN)ORACLE_FUSION)PING)PROOFPOINT_SENDMAIL_SENTRION)SAILPOINT_IAM)SALESFORCE)SENDMAIL)SENTINELONE_ALERT)SERVICENOW_AUDIT)SERVICENOW_CMDB)SERVICENOW_SECURITY)SONIC_FIREWALL)STIX)TANIUM_THREAT_RESPONSE)THINKST_CANARY)THREATCONNECT_IOC_V3)THREATLOCKER)VARONIS)VMWARE_ESX)WINDOWS_DNS)WINEVTLOG)WINEVTLOG_XML)WIZ_IO)WORKSPACE_ACTIVITY)ZOOM_OPERATION_LOGS)The following log types were added without a default parser. Each parser is listed by product name and log_type value, where applicable.
ADOBE_EXPERIENCE_PLATFORM)AUDIOCODES_SBC)AZURE_GATEWAY_CONTAINERS)AZURE_LOGIC_APPS)AZURE_NATGW_FLOW)BROADCOM_DX_NETOPS_SPECTRUM)CARTO_ACTIVITY)CLAUDE_CODE_OBSERVABILITY)CYBLE_ASM)CYBLE_BIP)DARKWEB_IQ)ELLIO_THREAT_INTEL)EXEON_NDR)GRAVITEE)KASPERSKY_ANTI_TARGETED_ATTACK)MICROSOFT_COPILOT_INTERACTION)OSTTRA_MARKITWIRE)PROOFPOINT_ADAPTIVE_EMAIL_SECURITY)SECOMEA_GATEMANAGER)TRENDMICRO_VISION_ONE_RISK_EVENT)TXONE_EDGEIPS)VECTRA_RUX)ZOHO_CRM)View prebuilt parser version content
You can now view the prebuilt parser preview version content even if you are using a custom parser for the same log type. Although the prebuilt parser version is inactive, you can still see the content of the new preview version for this parser.
Active Directory: Version 44.0
Added optional Connection Timeout and Receive Timeout parameters to
configure network connectivity limits in the following action:
Anomali ThreatStream: Version 18.0
Updated API output handling in the following action:
Google Threat Intelligence: Version 20.0
Added support for CHILDHASH and PARENTHASH entity types in the following
action:
Added Entity Type Filter parameter to allow configuring entity types for
notifications in the following connector:
Microsoft 365 Defender: Version 28.0
Updated case syncing logic in the following action:
Updated alert processing logic in the following connector:
Siemplify: Version 112.0
Added Update Enabled Connectors Only filtering option in the following job:
Vertex AI: Version 8.0
Google SecOps has updated the list of supported default parsers. Parsers are updated gradually, so it might take one to four days before you see the changes reflected in your region.
The following supported default parsers have been updated. Each parser is listed by product name and log_type value, where applicable. This list includes both released default parsers and pending parser updates.
AIRLOCK_DIGITAL)AIX_SYSTEM)AKAMAI_DATASTREAM_2)AKAMAI_SIEM_CONNECTOR)APACHE)ARCSIGHT_CEF)ARMIS_ALERTS)ARUBA_SWITCH)ATLASSIAN_AUDIT)AUDITD)AVAYA_AURA)AWS_CLOUDTRAIL)AWS_CLOUDWATCH)AWS_CONTROL_TOWER)AZURE_ACTIVITY)AZURE_AD)AZURE_AD_CONTEXT)AZURE_GATEWAY)AZURE_KEYVAULT_AUDIT)AZURE_RESOURCE_LOGS)BLUECOAT_WEBPROXY)BOMGAR)CATO_NETWORKS)CHECKPOINT_FIREWALL)CHECKPOINT_HARMONY)CHROME_MANAGEMENT)CHROMEOS_XDR)CISCO_ASA_FIREWALL)CISCO_EMAIL_SECURITY)CISCO_FIREPOWER_FIREWALL)CISCO_FIRESIGHT)CISCO_ISE)CISCO_ROUTER)CISCO_SWITCH)CISCO_UCM)CLAROTY_XDOME)CLAUDE_COMPLIANCE_LOGS)CLEARPASS)CLOUDFLARE)CORTEX_XDR)CS_EDR)DARKTRACE)EFFICIENTIP_DDI)F5_ASM)F5_BIGIP_LTM)FASTLY_CDN)FIREEYE_EMPS)FIREEYE_HX)FIREEYE_NX)FORCEPOINT_WEBPROXY)FORTINET_FIREWALL)FORTINET_FORTIANALYZER)FORTINET_FORTICLIENT)FORTINET_SWITCH)GCP_CLOUDAUDIT)GCP_SECURITYCENTER_EXTERNAL_EXPOSURE)GITLAB)GTI_IOC)GUARDDUTY)HUAWEI_SWITCH)IBM_SAM)IIS)ILLUMIO_CORE)IMPERVA_SECURESPHERE)INFOBLOX)INFOBLOX_DHCP)JAMF_PRO_CONTEXT)LOOKOUT_MOBILE_ENDPOINT_SECURITY)MACOS)MCAFEE_IPS)MICROFOCUS_IMANAGER)MICROSOFT_DEFENDER_ENDPOINT)MICROSOFT_DEFENDER_MAIL)MICROSOFT_GRAPH_ALERT)MICROSOFT_SENTINEL)MICROSOFT_SQL)MIMECAST_URL_LOGS)MISP_IOC)NETAPP_ONTAP)NETSKOPE_ALERT_V2)NIX_SYSTEM)OFFICE_365)OKTA)ONAPSIS)OPEN_VPN)ORACLE_FUSION)PING)PROOFPOINT_SENDMAIL_SENTRION)SAILPOINT_IAM)SALESFORCE)SENDMAIL)SENTINELONE_ALERT)SERVICENOW_AUDIT)SERVICENOW_CMDB)SERVICENOW_SECURITY)SONIC_FIREWALL)STIX)TANIUM_THREAT_RESPONSE)THINKST_CANARY)THREATCONNECT_IOC_V3)THREATLOCKER)VARONIS)VMWARE_ESX)WINDOWS_DNS)WINEVTLOG)WINEVTLOG_XML)WIZ_IO)WORKSPACE_ACTIVITY)ZOOM_OPERATION_LOGS)The following log types were added without a default parser. Each parser is listed by product name and log_type value, where applicable.
ADOBE_EXPERIENCE_PLATFORM)AUDIOCODES_SBC)AZURE_GATEWAY_CONTAINERS)AZURE_LOGIC_APPS)AZURE_NATGW_FLOW)BROADCOM_DX_NETOPS_SPECTRUM)CARTO_ACTIVITY)CLAUDE_CODE_OBSERVABILITY)CYBLE_ASM)CYBLE_BIP)DARKWEB_IQ)ELLIO_THREAT_INTEL)EXEON_NDR)GRAVITEE)KASPERSKY_ANTI_TARGETED_ATTACK)MICROSOFT_COPILOT_INTERACTION)OSTTRA_MARKITWIRE)PROOFPOINT_ADAPTIVE_EMAIL_SECURITY)SECOMEA_GATEMANAGER)TRENDMICRO_VISION_ONE_RISK_EVENT)TXONE_EDGEIPS)VECTRA_RUX)ZOHO_CRM)View prebuilt parser version content
You can now view the prebuilt parser preview version content even if you are using a custom parser for the same log type. Although the prebuilt parser version is inactive, you can still see the content of the new preview version for this parser.
Airflow 3.2.2 is available in Managed Airflow (Gen 3).
(Airflow 3.2.2) The
Multi-Team
Airflow feature isn't available. The [core]multi_team Airflow configuration
option is set to False and it isn't possible to override it.
(Airflow 3.2.2) Backported
#69877 to restore the ability
to deliver failure and retry alerts through a pluggable email backend
(configured through the [email]email_backend Airflow configuration option).
(Managed Airflow Gen 3 with Airflow 2) Default triggerer resources are changing to 1 vCPU and 2 GB memory to match Airflow 3 defaults. This change is available in the Google Cloud CLI, Terraform, and Cloud Composer API and is gradually rolling out in the Google Cloud console.
A correct error message is now generated when an environment creation request fails because of malformed network and subnetwork identifiers.
(Available without upgrading) The correct default task priority weight of 1
is now shown for tasks in the Google Cloud console.
New Airflow builds are available in Managed Airflow (Gen 3):
New images are available in Managed Airflow (Gen 2):
The following Managed Airflow versions and builds have reached their end of support period: composer-3-airflow-2.10.5-build.10, composer-3-airflow-2.9.3-build.30, composer-2.13.8-airflow-2.9.3, and composer-2.13.8-airflow-2.10.5.
The Agent overview dashboard is available on the Google Cloud Fraud Defense home page. This dashboard helps you monitor and analyze automated agent traffic on your site by distinguishing verified agents from suspected agents.
For more information, see Monitor agent traffic.
]]>IAM group authentication for AlloyDB is available in Preview for new clusters running PostgreSQL 15 and later. This feature simplifies database user management by allowing access management at the group level, where group members inherit database roles and permissions. To use this feature, enable the alloydb.iam_authentication and alloydb.iam_group_authentication database flags.
For more information, see IAM group authentication and Manage IAM authentication.
Hyperdisk Balanced volumes on C4D instances have increased maximum throughput limits for these machine types:
c4d-*-96: 3,125 MiB/s (up from 2,800 MiB/s).c4d-*-192: 6,250 MiB/s (up from 4,800 MiB/s).c4d-*-384: 12,500 MiB/s (up from 10,000 MiB/s).For detailed performance limits, see Hyperdisk Balanced performance limits when attached to an instance.
Starting August 2026, Confidential VM instances using AMD SEV-SNP might have longer boot times and performance changes due to a guest kernel migration and security updates. This issue is expected to be resolved by November 2026. Confidential VM instances using AMD SEV or Intel TDX aren't affected.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.12.94 | v27.5.1 | v2.2.5 | See List |
Updated udev rule for protected_stateful_partition
Upgraded app-admin/fluent-bit to v4.2.7.
Fixed CVE-2026-29111 in sys-apps/systemd
Fixed CVE-2026-3644 in dev-lang/python
Fixed CVE-2026-40355 and CVE-2026-40356 in app-crypt/mit-krb5.
Fixed CVE-2026-53381 in the Linux kernel.
Fixed CVE-2026-53385 in the Linux kernel.
Fixed CVE-2026-53388 in the Linux kernel.
Fixed CVE-2026-53391 in the Linux kernel.
Fixed CVE-2026-53392 in the Linux kernel.
Fixed CVE-2026-53393 in the Linux kernel.
Fixed CVE-2026-53394 in the Linux kernel.
Fixed CVE-2026-53397 in the Linux kernel.
Fixed CVE-2026-53398 in the Linux kernel.
Fixed CVE-2026-53400 in the Linux kernel.
Fixed CVE-2026-59995, CVE-2026-59996, CVE-2026-59997, CVE-2026-59999, CVE-2026-60000, CVE-2026-60001, and CVE-2026-60002 in openssh.
Fixed CVE-2026-6019 in dev-lang/python
Fixed CVE-2026-63795 in the Linux kernel.
Fixed CVE-2026-63800 in the Linux kernel.
Fixed CVE-2026-63802 in the Linux kernel.
Fixed CVE-2026-63806 in the Linux kernel.
Fixed CVE-2026-63807 in the Linux kernel.
Fixed CVE-2026-63809 in the Linux kernel.
Fixed CVE-2026-63810 in the Linux kernel.
Fixed CVE-2026-63823 in the Linux kernel.
Fixed CVE-2026-63824 in the Linux kernel.
Fixed CVE-2026-63827 in the Linux kernel.
Fixed CVE-2026-63828 in the Linux kernel.
Fixed CVE-2026-63829 in the Linux kernel.
Fixed CVE-2026-63830 in the Linux kernel.
Fixed CVE-2026-63833 in the Linux kernel.
Fixed CVE-2026-64187 in the Linux kernel.
Fixed CVE-2026-64189 in the Linux kernel.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.12.94 | v27.5.1 | v2.1.9 | See List |
Fixed an important bug for xfs file system users.
Updated udev rule for protected_stateful_partition
Fixed CVE-2026-29111 in sys-apps/systemd
Fixed CVE-2026-3644 in dev-lang/python
Fixed CVE-2026-40355 and CVE-2026-40356 in app-crypt/mit-krb5.
Fixed CVE-2026-53381 in the Linux kernel.
Fixed CVE-2026-53385 in the Linux kernel.
Fixed CVE-2026-53388 in the Linux kernel.
Fixed CVE-2026-53391 in the Linux kernel.
Fixed CVE-2026-53392 in the Linux kernel.
Fixed CVE-2026-53393 in the Linux kernel.
Fixed CVE-2026-53394 in the Linux kernel.
Fixed CVE-2026-53397 in the Linux kernel.
Fixed CVE-2026-53398 in the Linux kernel.
Fixed CVE-2026-53400 in the Linux kernel.
Fixed CVE-2026-59995, CVE-2026-59996, CVE-2026-59997, CVE-2026-59999, CVE-2026-60000, CVE-2026-60001, and CVE-2026-60002 in openssh.
Fixed CVE-2026-6019 in dev-lang/python
Fixed CVE-2026-63795 in the Linux kernel.
Fixed CVE-2026-63800 in the Linux kernel.
Fixed CVE-2026-63802 in the Linux kernel.
Fixed CVE-2026-63806 in the Linux kernel.
Fixed CVE-2026-63807 in the Linux kernel.
Fixed CVE-2026-63809 in the Linux kernel.
Fixed CVE-2026-63810 in the Linux kernel.
Fixed CVE-2026-63823 in the Linux kernel.
Fixed CVE-2026-63824 in the Linux kernel.
Fixed CVE-2026-63827 in the Linux kernel.
Fixed CVE-2026-63828 in the Linux kernel.
Fixed CVE-2026-63829 in the Linux kernel.
Fixed CVE-2026-63830 in the Linux kernel.
Fixed CVE-2026-63833 in the Linux kernel.
Fixed CVE-2026-64187 in the Linux kernel.
Fixed CVE-2026-64189 in the Linux kernel.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.6.143 | v24.0.9 | v1.7.34 | See List |
Upgraded net-fs/cifs-utils to v7.7, Upgraded sys-libs/talloc to v2.4.4-r1.
Fixed CVE-2026-29111 in sys-apps/systemd
Fixed CVE-2026-40355 and CVE-2026-40356 in app-crypt/mit-krb5.
Fixed CVE-2026-53381 in the Linux kernel.
Fixed CVE-2026-53385 in the Linux kernel.
Fixed CVE-2026-53388 in the Linux kernel.
Fixed CVE-2026-53391 in the Linux kernel.
Fixed CVE-2026-53397 in the Linux kernel.
Fixed CVE-2026-53398 in the Linux kernel.
Fixed CVE-2026-6019 in dev-lang/python
Fixed CVE-2026-63794 in the Linux kernel.
Fixed CVE-2026-63795 in the Linux kernel.
Fixed CVE-2026-63800 in the Linux kernel.
Fixed CVE-2026-63802 in the Linux kernel.
Fixed CVE-2026-63807 in the Linux kernel.
Fixed CVE-2026-63809 in the Linux kernel.
Fixed CVE-2026-63823 in the Linux kernel.
Fixed CVE-2026-63824 in the Linux kernel.
Fixed CVE-2026-63827 in the Linux kernel.
Fixed CVE-2026-63828 in the Linux kernel.
Fixed CVE-2026-63830 in the Linux kernel.
Gemini Enterprise: Microsoft Teams federated data store is generally available
The Microsoft Teams federated data store is generally available (GA) in Gemini Enterprise. Connect Microsoft Teams to query channels, chats, teams, and messages, and execute supported actions directly from the assistant.
For more information, see Connect Microsoft Teams.
GKE now supports opting out of the default kubernetes.io/arch=arm64:NoSchedule
taint on Arm nodes in Standard node pools and in custom ComputeClasses. To opt
out of the default taint, set the --node-architecture-taint-behavior gcloud
CLI flag to NONE for a node pool or set the
taintConfig.architectureTaintBehavior field to NONE for a ComputeClass. By
configuring this behavior, you allow workloads that lack explicit Arm
tolerations to be scheduled on Arm-based machine families (such as N4A and C4A).
This is useful for running multi-architecture workloads or simplifying
scheduling in mixed-mode clusters. For more information, see
Configure the default Arm architecture taint.
GKE Gateway and Inference Gateway now support Cross-Origin Resource Sharing
(CORS). You can configure a CORS filter directly on an HTTPRoute resource by
using the portable syntax standardized by
Gateway API.
This feature is available in Preview in GKE version 1.35 and later for the
following GatewayClasses:
gke-l7-rilbgke-l7-regional-external-managedgke-l7-global-external-managedFor more information, see Configure Cross-Origin Resource Sharing.
[Spotlight Feature] Data RBAC for first-party (1P) cases and alerts
Availability This feature is now available in public preview for all regions.
Google SecOps now supports data role-based access control (Data RBAC) for first-party (1P) SOAR cases and alerts. This feature automatically applies SIEM data access scopes to alerts and cases ingested using the Chronicle connector, ensuring analysts only see data they are authorized to access.
For more information, see the Release Note entry for July 6th.
Starting with Managed Service for Apache Spark image version 3.0,
clusters that are created without a specified machine type for a node are created with a
Flex VM
configuration for the node.
Version 1.2.0 of the Google SCC ITSM app and version 1.3.0 of the Google SCC SIR app have been released.
To reflect this update, the ServiceNow integration guide is updated with the following changes:
Added the following features:
Updated setup instructions for Java KeyStore certificates.
Added additional troubleshooting steps for maximum execution time exceeded errors, data collection issues, and ECC Queue timeout errors.
For more information, see Sending Security Command Center data to ServiceNow.
Spanner supports creating tables without defining primary keys.
When you create a table without a primary key, Spanner creates a
hidden rowid column that serves as the primary key. For more information, see
Create a table without defining a primary key.
AlloyDB write endpoints are now available in Preview. Write endpoints simplify database connection management by providing a stable domain name service (DNS) name for your applications, decoupling them from instance IP addresses. During disaster recovery switchovers or failovers, AlloyDB automatically updates the endpoint to point to the new primary instance.
For more information, see Manage database connections with write endpoints.
AlloyDB now supports cross-region failover in Preview. You can optionally enable cross-region failover to automate the recreation of the original primary when a secondary cluster is promoted while maintaining your replication topology without requiring manual cluster deletion or re-creation. Cross-region failover is supported only for topologies with a single primary cluster and a single secondary cluster.
For more information, see About cross-region replication and Work with cross-region replication.
External search with AlloyDB now supports OpenSearch in Preview.
You can use the external_search_fdw extension to connect to an OpenSearch cluster and query its data directly from your database.
Upgraded bundled Helm version from v3.20.2 to v3.21.1 to pick up vulnerability fixes. To understand the changes in each release, review the changelogs.
Addressed multiple Common Vulnerabilities and Exposures (CVEs) by updating dependencies.
On July 27th, 2026, we released an updated version of Apigee (1-18-0-apigee-2).
SanitizeUserPrompt and SanitizeModelResponse policies now support function calling
The SanitizeUserPrompt and SanitizeModelResponse policies now handle LLM function calling (tool use) flows. Two new optional XML elements extract function calling content for Model Armor sanitization:
<FunctionResponseSource> in SanitizeUserPrompt: extracts function/tool
response data from user turns. Used as a fallback when <UserPromptSource>
returns empty on function calling turns.<FunctionCallSource> in SanitizeModelResponse: extracts function call
arguments from model responses. Used as a fallback when <LLMResponseSource>
returns empty.Default templates target the Gemini format and require no additional configuration. For OpenAI and Anthropic formats, configure custom JSONPath templates. For more information, see FunctionResponseSource and FunctionCallSource.
| Bug ID | Description |
|---|---|
| 534852923 | Security fix for Apigee. Fixed a security issue in the Java Callout policy. |
| N/A | Security fix for Apigee infrastructure. |
| Bug ID | Description |
|---|---|
| N/A | Updates to infrastructure and libraries. |
Support for the Node.js 26 runtime is in Preview.
Support for the Node.js 26 runtime is in Preview.
You can discover commercial BigQuery sharing listings on Google Cloud Marketplace with the Marketplace filter. For more information, see Subscribe to a Cloud Marketplace-integrated listing. This feature is generally available (GA).
You can use the APPENDS and CHANGES change history functions
to view the rows that were appended to or changed in a table during a given time
range. This feature is
generally available
(GA).
You can use the Google-developed Open Database Connectivity (ODBC) driver for BigQuery to connect your applications to BigQuery. This feature is generally available (GA).
The feature formerly known as the legacy tabledata.insertAll method is now
called the
Storage Write API (REST). The
feature formerly known as the Storage Write API is now called the
Storage Write API (gRPC).
BigQuery dataset insights is generally available (GA).
BigQuery dataset insights helps you discover and visualize relationships between tables and automatically generate cross-table queries. You can run and publish these insights to Knowledge Catalog for agentic grounding use cases, or generate them on demand without publishing for quick, ad hoc dataset exploration.
The Bigtable remote MCP server supports the Bigtable Data API, which
provides the execute_sql tool that you can use to query Bigtable data using
natural language prompts. This feature is
generally available (GA).
For more information, see
Use the Bigtable remote MCP server.
Spend cap budgets are now available for a limited set of services (Preview)
Available in Preview for eligible services, you can now configure a spend cap budget to automatically pause usage when your spend exceeds the budget amount you set.
Spend caps are a cost control mechanism. A spend cap is enforced when usage costs exceed your budget target amount. When enforced, any new request to the eligible services, within the specified project, are paused and no further usage costs are accrued until you manually lift the spend cap.
Spend caps typically use estimated costs to trigger the alerts and caps, enforcing a cap much faster than the actual costs are processed and appear on billing reports. Even though faster than reports, the enforcement of spend caps isn't instant and any cost overages are billed as normal.
For more information about spend cap budgets, see:
Service load balancing policies (serviceLbPolicy) are now supported for
regional external Application Load Balancers and regional internal Application Load Balancers. This feature enables
advanced load balancing optimizations such as custom load balancing algorithms,
auto-capacity draining, failover thresholds, and the ability to designate
preferred backends for these load balancers.
For more information, see Advanced load balancing optimizations.
This feature is in Preview.
Enabling WildFire in an existing firewall endpoint can cause a temporary data plane outage. As a result, the WildFire feature is temporarily removed.
Support for the Node.js 26 runtime is in Preview.
Support for Budget spend caps to pause your Cloud Run workloads is in Preview.
Support for the Node.js 26 runtime is in Preview.
The maximum IOPS per GiB for Hyperdisk Balanced Storage Pools have increased from 4 IOPS per GiB. The new limits depend on the provisioning type:
For more information, see Limits for Hyperdisk Storage Pools.
Dataform deployments provide a centralized experience for creating and managing pipeline deployments connected to remote Git repositories. This feature is available in Preview.
Gemini Enterprise: Transparent thinking
Transparent thinking is generally available (GA). During chat interactions, the assistant shares its real-time reasoning and planning in the user interface before calling tools or data sources.
An expandable section displays the tool activity between the thinking phase and the final answer. This transparency delivers a faster time to first token (TTFT) and improves perceived latency without increasing total response time.
For more information, see Ask questions and view sources.
You can now create a Cloud SQL for PostgreSQL Source connector and a Generic PostgreSQL Source connector for Kafka Connect.
A Cloud SQL for PostgreSQL Source connector or Generic PostgreSQL Source connector is an instance of a Debezium PostgreSQL connector. It reads row-level changes from a PostgreSQL database and writes them to topics in a Managed Service for Apache Kafka cluster.
For more information, see Create a Cloud SQL for PostgreSQL Source connector, Create a Generic PostgreSQL Source connector, and Troubleshoot a PostgreSQL Source connector.
In GKE version 1.36 and later, GKE Dataplane V2 with NetworkPolicies supports up to 15,000 nodes per cluster, increased from the previous limit of 7,500 nodes. For clusters exceeding 5,000 nodes, contact Cloud Customer Care to request a quota increase. For more information, see Cluster size limits and requirements.
In version 1.36.2-gke.1498000 and later, GKE supports mixed-protocol Services of type LoadBalancer in general availability (GA). Mixed-protocol Services let both external (NetLB) and internal (ILB) passthrough Network Load Balancers handle simultaneous TCP and UDP traffic on a single IP address across IPv4, IPv6, and dual-stack environments.
The general availability (GA) stage of mixed-protocol Services of type LoadBalancer fixes errors in traffic routing from stages prior to GA. This feature is in the GA stage in GKE version 1.36.2-gke.1498000 and later.
Version 20260716.00 of the guest agent
is now available for all supported operating systems. This version introduces
the following fixes:
systemctl start operations performed by the OS Login module no longer
leave behind zombie processes.Managed workload identities for Compute Engine are generally available.
For more information, see Configure managed workload identity authentication for Compute Engine.
Added support for Valkey version 9.1. As a result, you can now upgrade the version of your Memorystore for Valkey instance to 9.1. For more information, see About upgrading the Valkey version of an instance. This feature is available in Preview.
Automatic rotation of regional Cloud SQL database credentials in Secret Manager is available in Preview. This feature lets you automatically generate secure passwords, update target Cloud SQL database instances (PostgreSQL or SQL Server), and rotate secret versions on a configured schedule without custom Cloud Run functions.
For more information, see Automatic rotation of Cloud SQL secrets.
For the Security Command Center Premium tier, you can enable AI Protection at the project level.
Project-level activations include access to the AI security dashboard, AI threat detection, and AI vulnerability and misconfiguration findings.
Some features of AI Protection are only available for organization-level activations. For more information, see Configure AI Protection.
For the Security Command Center Premium tier, you can enable AI Protection at the project level.
Project-level activations include access to the AI security dashboard, AI threat detection, and AI vulnerability and misconfiguration findings.
For more information, see Configure AI Protection.
]]>Customizable schedules for multi-event rules
Customizable schedules for multi-event rules are available in public preview. You can customize rule execution schedules on the Rule schedule tab to specify a first-run delay offset that accounts for data ingestion latency. The system also performs automated background true-up runs to catch late-arriving logs and process metadata enrichment without requiring manual system interventions. This gives you precise control over detection evaluation timing, reduces false negatives without missing detections, and promotes alert accuracy.
To view or modify rule schedules using custom Identity and Access Management (IAM) roles, update your IAM permissions to include the following:
chronicle.ruleDeployments.update to update individual rule schedules using the API.chronicle.rules.modifyRules to modify rule schedules using the web interface or in batch using the API.If you use predefined IAM roles, such as Chronicle API Admin (roles/chronicle.admin) or Chronicle API Editor (roles/chronicle.editor), these permissions are included automatically.
[Spotlight Feature] Investigation and case management experience
This feature is in public preview. Google SecOps now includes a revamped Investigation Management experience that supports tracking raw UDM events and detections alongside alerts to accommodate new investigation types (such as retrohunt and threat hunt) and higher investigation volumes in cases. You can navigate your case queue using customizable table views, side-drawer previews, and integrated UDM Search workflows. For more information, see Investigation and case management overview.
This preview is currently supported only for single-SIEM deployments (instances where a single Google SecOps SIEM instance ingests data into SOAR) and does not support federated or MSSP environments.
Additional enhancements include:
Release 6.3.95 is being rolled out to the first phase of regions as listed here.
This release contains internal and customer bug fixes.
]]>Release 6.3.94 is now available for all regions.
]]>General availability (GA) launch of Model Context Protocol (MCP) in API hub
The API hub MCP server is Generally Available (GA). This release enables seamless integration between your AI agents and API hub, allowing your applications to discover, query, and manage your API ecosystem using natural language.
What's new in GA
apihub.googleapis.com/mcp), in addition to the supported regional endpoints. For a list of supported regions, see the API hub MCP reference.apihub.readonly and apihub.readwrite) for more secure access.For configuration details and a complete list of available tools, see API hub MCP reference.
Configure and deploy MCP servers with API hub RPC
API hub now includes a new RPC, ConfigureAndDeployServer, which enables the configuration and deployment of Model Context Protocol (MCP) servers directly to an Apigee runtime.
For more information, see Manage MCP proxies and the API hub reference.
On July 24, 2026 we released an updated version of the Apigee hybrid software, v1.16.8.
| Bug ID | Description |
|---|---|
| 493354568 | Fixed an issue where component-specific nodeSelector configurations are ignored in Helm charts. |
Runtime rollout strategy configuration
In this release, you can configure the rollout strategy used when updating runtime (message processor) ReplicaSets by setting the runtime.release.strategy property (with options rolling, scale-down-first, or none) or per-environment with envs[].components.runtime.release.strategy in your overrides configuration file. The property defaults to rolling.
Various security and CVE fixes are included in this release.
Connector repositories act as proxies for upstream sources. All requests to the repository are proxied to the upstream source and no artifacts are cached in Artifact Registry. This configuration allows for full auditability of upstream sources and supports cases where third-party policies prevent artifact caching. For more information, see Connector repositories overview.
You can now use Model Context Protocol (MCP) servers with Backup and DR Service to connect AI assistants and applications—such as the Gemini CLI, ChatGPT, or Claude—with your backup environment. Using either the local or remote MCP server, you can use natural language prompts to perform and automate Backup and DR tasks, such as creating backup plans, triggering on-demand backups, and managing backup vaults. Comprehensive reference documentation is also available with detailed specifications, input and output schemas, and sample invocation commands for all available MCP tools.
For more information, see the following:
As of July 30, 2026, support for Model Context Protocol (MCP) servers and tools will be shut down. You will not be able to retrieve, list, enable, and disable MCP servers and tools using the Cloud API Registry API. For more information, see Feature deprecations.
Early signals for AI workloads
For AI workloads (such as Gemini API and Vertex AI), you can now view early anomalies. Early anomalies use near real-time cost estimates to provide daily, service-level insights before finalized billing occurs. You can view these alerts on the By service (Early signals) tab on the Anomalies dashboard in the Google Cloud console. User-configured thresholds do not apply to early anomalies.
For more information, see View early anomalies for AI workloads.
Cloud SQL for PostgreSQL now supports logical replication using failover slot which you can use with advanced disaster recovery (DR) switchover and replica failover operations to ensure business continuity.
For more information, see Advanced disaster recovery (DR) with logical failover slot.
The Envoy Compressor Filter is now GA in the stable release channel.
You can observe real-time virtual machine (VM) distribution across zones, machine types, and instance states in your managed instance groups (MIGs) by using the GCE MIG Instance Distribution Monitoring dashboard in Cloud Monitoring. When your group uses location flexibility across zones, instance flexibility across machine types, or both, this visibility helps you monitor capacity allocation and diagnose runtime fallback behavior. For more information, see Monitor instance distribution in MIGs.
Gemini Enterprise: Gemini 3.6 Flash in US multi-region
If your project is on the allowlist, you can use Gemini 3.6 Flash in the US
multi-region (us) with data residency at-rest (DRZ) and machine learning
processing (MLP).
To request access to Gemini 3.6 Flash in the US multi-region, contact your Google account team.
For more information, see Data residency for Gemini Enterprise Standard and Plus Editions and Gemini Notebook Enterprise
Gemini Enterprise: Create and edit documents and slides in Canvas (GA)
The Canvas assistant is generally available within the Gemini Enterprise web app. Canvas is a dedicated, interactive tool thats allows you to create and edit AI-generated documents and presentations directly from your chats. You can then export these to Google Workspace, Microsoft Office formats, and PDF.
For more information, see Create and edit documents and slides in Canvas.
A Gemini Enterprise app administrator must turn on the Enable canvas toggle in the web app feature management settings to let users use it. For more information about feature controls, see Manage features on the web app.
Anthropic's Claude Opus 5
Claude Opus 5 is available in Model Garden.
Google Cloud CCaaS 5.0
We've released version 5.0 of Google Cloud CCaaS.
The timing of the update to your instance depends on the deployment schedule that you have chosen. For more information, see Deployment schedules.
Disposition timing
You can now configure CCAI Platform so agents can attribute wrap-up time, disposition code, and notes to a previous interaction. You can also let agents modify the disposition code and notes of previously completed sessions.
Administrators: In the Settings > Operation Management > Wrap-up pane, a new Manual Wrap-up section is available.
User experience change: If configured, a new Previous Sessions list appears in the agent adapter.
For more information, see Configure disposition timing.
API direct access point for chat
The API direct access point (DAP) for chat lets you automatically route incoming chat sessions to a queue based on a response from an external API endpoint that you configure. This eliminates the need for end-users to select from a queue menu.
By default, this capability is inactive. Contact Google Cloud Support to turn it on for your instance.
For more information, see API direct access point for chat.
This release addresses the following issues:
Fixed an issue where emails were stuck in a Transferring state when moved between queues.
Fixed an issue where end-users didn't receive messages from agents during web chats.
Fixed an issue where calls or chats remained in a queue without being offered to available agents.
Fixed an issue where voice calls were prematurely moved from an agent's queue during a multi-group cascade.
Fixed an issue where the disposition panel didn't appear after a call ended, leaving agents unable to change their status without signing out and signing in.
Fixed an issue where end-users were incorrectly assigned to teams and skills they weren't originally part of during bulk CSV imports.
Fixed an issue where interaction transcripts for calls in non-English languages were incorrect.
Fixed an issue in Salesforce integrations where incoming chat audio and desktop notifications didn't play for agents using the embedded CCAI Platform widget.
Fixed an issue where the inactive chat dismissal timer didn't reset after a chat was transferred from a virtual agent to a human agent.
Fixed an issue where voicemails disappeared from the queue immediately after being opened.
Fixed an issue where the reporting dashboard incorrectly displayed call and agent status during a cold transfer to another queue.
Improved rendering performance in the agent desktop mini chat adapter.
Fixed an issue where the storage path for screen recordings didn't align with the folder structure displayed in the user interface.
Fixed an issue where missing public files were incorrectly cached by the CDN for up to seven days.
Fixed an issue where agents were automatically redirected to the Closed inbox view after changing an interaction status to Closed.
Fixed an issue where the You cannot log out when in a chat notification was truncated in the chat adapter.
Fixed an issue with Salesforce integrations where rapid, concurrent data requests caused information to be lost.
Fixed an issue where duplicate call recording links were posted to Zendesk tickets for multi-segment calls.
Fixed an issue with Salesforce integrations where the UI retained settings from a previous Salesforce organization after switching to a new organization.
Fixed an issue where duplicate customer satisfaction surveys were submitted and recorded for a single live chat session.
Fixed an issue where the message field in the chat adapter was inactive when an agent accepted a new chat.
Fixed an issue where task virtual agents were incorrectly identified as Nobody when joining a conversation after a transfer from a human agent.
Fixed an issue that occurred when a chat entered a queue and the greeting message was sent before an agent was assigned. In these cases, the associated push notification crashed and logged an error, producing excessive noise in logs.
Fixed an issue where the agent's final message in a chat session appeared after This chat is ended in the chat adapter and the CRM transcript.
Fixed an issue where agents using instances without a CRM configuration received a No Account Detected warning when making outbound calls.
Fixed an issue where notification chimes played after an agent had connected to an active call.
Fixed an issue where the country code list didn't automatically update the
country flag when a phone number was entered without the + prefix.
Fixed an issue where calls to an agent's personal queue didn't break through if the agent was also assigned to an inbound queue with breakthrough disabled.
Fixed an issue where outbound calls that were transferred to a queue didn't adhere to the queue's deltacast configuration.
Fixed an issue where the agent desktop became unstable or didn't load.
Fixed an issue where wrap-up time was incorrectly reported when agents exceeded the configured wrap-up time.
Fixed an issue where the IVR queues dashboard didn't load for instances with a large volume of queues.
Fixed an issue where calls to the manager/api/v1/agent_activity_logs
endpoint timed out when sort_direction was turned off.
You can now use the Cloud Commerce Producer API to programmatically create, manage, and publish private offers. The API lets you automate your private offer workflows, including configuring custom pricing models, attaching EULA or SOW documents, and determining which active offer to amend.
For more information, see Create and manage private offers using the API and Determine which offer to amend.
A release note published on May 6, 2025 stated that each Keepalived instance
virtual router redundancy protocol (VRRP) configuration is configured with a
nopreempt flag to avoid elections when a non-master instance is restarted.
The nopreempt flag was removed in release 1.32.200 and later. For more
information, see Control plane VIP isn't moved when HAProxy is unavailable.
GKE cluster versions have been updated.
New versions available for upgrades and new clusters.
The following versions are now available for new GKE clusters, and for manual control plane upgrades and node upgrades for existing clusters. For more information about versioning and upgrades, see GKE versioning and support and About GKE cluster upgrades.
This release includes new GKE versions that use updated Container-Optimized OS images. These updated images are cumulative, incorporating security fixes from all Container-Optimized OS versions released since the previous GKE release.
To identify the specific vulnerabilities that were resolved in each updated Container-Optimized OS image, see the Security release notes for that image. The following table includes links to the release notes for each updated Container-Optimized OS image:
| GKE version | Container-Optimized OS version | Details |
|---|---|---|
| 1.32.13-gke.2137000 | cos-117-18613-675-2 | cos-117-18613-675-2 release notes |
| 1.34.9-gke.1610000 | cos-125-19216-532-3 | cos-125-19216-532-3 release notes |
| 1.36.2-gke.2064000 | cos-129-19506-299-3 | cos-129-19506-299-3 release notes |
Governance workflows let you set up automated controls for data product access management by providing a request-review mechanism. This feature is available in preview.
For more information, see About governance workflows.
Starting in September, 2026, Airflow 2.10.5 will no longer be included in new Managed Airflow images and builds. This change will not affect existing images and builds.
Starting in September 2026, we are changing the version support policy for Managed Airflow (Gen 2) to align it with the Managed Airflow (Gen 3) policy. The changes will affect Airflow 2 versions that we release:
New SAP certification for operating system: SLES 16 for SAP
For use with SAP HANA and SAP NetWeaver on Google Cloud, SAP has certified the operating system SUSE Linux Enterprise Server (SLES) 16 for SAP.
For more information about SAP-certified operating systems, see:
As of July 30, 2026, support for managing Model Context Protocol (MCP) server enablement and consumer policies using the Service Usage v2beta API will be shut down.
For more information, see Feature deprecations.
Preview: Policy Engine, Universal keys, and challenge policies are available in Preview for Google Cloud Fraud Defense.
For more information about the types of challenges that are available, see Challenge types.
]]>Agent skill governance is available in Preview.
Agent Registry supports standalone skill governance, helping enable secure, enterprise-level management and governance of standalone skills for AI agents. You can register skill resources, upload and validate ZIP payload packages, track version history through immutable skill revisions, and review verified skill publishers.
This release includes the following features:
Skill resources, lifecycle states, and version snapshots (SkillRevision).For more information, see Register skills and Manage skills.
Transparent query forwarding is now available in Preview for clusters compatible with PostgreSQL 17 and 18. With this feature, the primary node in a cluster intercepts read-only queries and selectively forwards them to read pool instances while maintaining read-your-writes consistency.
For more information, see Optimize resources and isolate read queries with transparent query forwarding.
On July 23, 2026 we released an updated version of the Apigee hybrid software, v1.14.7.
Runtime rollout strategy configuration
In this release, you can configure the rollout strategy used when updating runtime (message processor) ReplicaSets by setting the runtime.release.strategy property (with options rolling, scale-down-first, or none) or per-environment with envs[].components.runtime.release.strategy in your overrides configuration file. The property defaults to rolling.
Various security and CVE fixes are included in this release.
An updated version of the Simba ODBC driver for BigQuery is now available.
Model Context Protocol (MCP) integration is available in Preview. This built-in integration lets LLM-powered agents securely retrieve data using standard MCP tools (search_cloud_locations and list_cloud_locations). For more information, see Use the Model Context Protocol (MCP) with Cloud Location Finder.
Cloud Router support for named sets for BGP route policies is now generally available. For more information, see BGP route policies overview.
Agent Search: Decrease thresholds for configurable pricing
You can decrease the storage size and queries per minute (QPM) subscription thresholds for configurable pricing. Previously, you could only increase these thresholds.
Decreased thresholds take effect at the start of the next billing cycle.
For more information, see Modify subscription thresholds. This feature is generally available (GA).
]]>Agent Identity is available in Preview for Access Approval.
Privileged Access Manager is available in Preview.
Agent Identity is available in Preview for Access Transparency.
Privileged Access Manager is available in Preview.
Various bug fixes and minor product enhancements.
Gemini Enterprise: Support for unauthenticated Custom MCP Server data stores (Preview)
When setting up a Custom MCP Server data store, you can select No authentication if your Model Context Protocol (MCP) server doesn't require authentication.
Creating a custom MCP Server data store is in Public Preview. For more information, see Set up a custom MCP server.
Google Distributed Cloud (software only) for VMware 1.35.300-gke.87 is now available for download. To upgrade, see Upgrade clusters. Google Distributed Cloud 1.35.300-gke.87 runs on Kubernetes v1.35.3-gke.400.
If you are using a third-party storage vendor, check the Google Distributed Cloud-ready storage partners document to make sure the storage vendor has already passed the qualification for this release.
After a release, it takes approximately 7 to 14 days for the version to become available for use with GKE On-Prem API clients: the Google Cloud console, the gcloud CLI, and Terraform.
Google Distributed Cloud (software only) for VMware 1.34.700-gke.93 is now available for download. To upgrade, see Upgrade clusters. Google Distributed Cloud 1.34.700-gke.93 runs on Kubernetes v1.34.7-gke.200.
If you use a third-party storage vendor, check the listing of our previously-qualified storage partners.
After a release, it takes approximately 7 to 14 days for the version to become available for use with GKE On-Prem API clients: the Google Cloud console, the gcloud CLI, and Terraform.
The following issues were fixed in 1.34.700-gke.93:
spec.selector fields on existing ANG resources. The upgrade operator now preserves existing label selectors during reconciliation so that V1 to V2 cluster migrations complete successfully.
The following issues were fixed in 1.35.300-gke.87:
spec.selector fields on existing ANG resources. The upgrade operator now preserves existing label selectors during reconciliation so that V1 to V2 cluster migrations complete successfully.
Google Distributed Cloud (software only) for bare metal 1.34.700-gke.93 is now available for download. To upgrade, see Upgrade clusters. Google Distributed Cloud for bare metal 1.34.700-gke.93 runs on Kubernetes v1.34.7-gke.200.
After a release, it takes approximately 7 to 14 days for the version to become available for installations or upgrades with the GKE On-Prem API clients: the Google Cloud console, the gcloud CLI, and Terraform.
If you use a third-party storage vendor, check the listing of our previously-qualified storage partners.
The following change was added in 1.34.700-gke.93:
csi-snapshot-validation-webhook component. Upstream Kubernetes validation is now handled natively via Common Expression Language (CEL) rules within the deployed Custom Resource Definitions (CRDs). For more information, see Volume snapshots.
The following issues were fixed in 1.34.700-gke.93:
Wiz: Version 14.0
Added the following action:
SentinelOne Singularity Operations Center: Version 1.0
Proofpoint Email Protection: Version 10.0
Added the following action:
Azure Monitor: Version 5.0
QRadar: Version 69.0
Updated timestamp filtering to use last_persisted_time for tracking
modifications in the following connector:
Jira: Version 60.0
Added support for the Created Before date filter and Custom JQL query
parameter in the following action:
Google Chronicle: Version 90.0
Updated handling of Wiz Defend detections and ontology mapping in the following connector:
A Cross-Site Scripting (XSS) vulnerability was discovered in Looker. An attacker could craft a malicious URL that, when opened by a Looker administrator, would allow the attacker to execute arbitrary scripts on their behalf and potentially compromise the administrator account.
Both Looker-hosted and self-hosted instances were found to be vulnerable.
This issue has already been mitigated for Looker-hosted instances.
What should I do?
For Looker-hosted instances, no action is required.
For self-hosted Looker instances, update your Looker instances as soon as possible. This vulnerability has been patched in all supported versions of Looker for self-hosted instances. The following versions have all been updated to fix this vulnerability:
For more information, see CVE-2026-15810.
]]>You can now launch clusters with the following Kubernetes versions. Click on the following links to see the release notes associated with these patches:
You can now launch clusters with the following Kubernetes versions. Click on the following links to see the release notes associated with these patches:
You can now launch clusters with the following Kubernetes versions. Click on the following links to see the release notes associated with these patches:
To provide long-term security and address threats from future quantum computers,
Binary Authorization supports keys that use post-quantum cryptography (PQC)
algorithms. These algorithms, such as ML-DSA-65 (Dilithium3), are standardized
to be resistant to attacks from both classical and quantum computers. To learn how
to generate a PQC key pair and create an attestor, see
Create post-quantum cryptography (PQC) keys.
The Envoy Lua Filter is now available as a preview feature in the stable release channel.
Cloud Tasks support for the following is available in Preview:
Generally available: You can configure a regional managed instance group (MIG) to allow a VM repair in an alternate zone when the MIG can't repair the VM in its original zone. Repairing a VM in an alternate zone can help to improve your application's resiliency and resource obtainability. For more information, see Repair a VM in an alternate zone.
Various bug fixes and minor product enhancements.
Gemini Enterprise: Gemini 3.6 Flash available in the Global region
Gemini 3.6 Flash is available in the global region. To make Gemini 3.6 Flash
available to users in the Gemini Enterprise app, administrators must turn on
the Gemini 3.6 Flash feature toggle.
Gemini 3.6 Flash is also available in Agent Designer workflow agents. Updates take up to a day to appear in workflow agents.
For more information, see:
Gemini Enterprise: Gemini 3.5 Flash removal in the Global region
Gemini 3.5 Flash will be removed as a model from the global region in the
Gemini Enterprise app on August 4, 2026.
For more information, see Manage features on the web app.
Gemini 3.6 Flash and 3.5 Flash-Lite are generally available (GA)
Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now generally available (GA) and available for production use. These models are designed to improve upon their predecessors' capabilities, including improved token usage and improved document understanding. See the linked model information pages for more information.
This release includes some potentially breaking changes from previous Flash and Flash-Lite models:
Model are no longer supported and will return an error:
"type": "model_output" will fail."role": "model" will fail.CodeMender available in Public Preview
CodeMender is now available in Preview. CodeMender is an AI code security agent that can find, verify, and fix deep vulnerabilities in your codebase. CodeMender wraps a fine-tuned harness around a large language model (LLM), using Google DeepMind-engineered prompts, skills, and orchestration logic to turn the model into an agentic system specialized in code security.
Key capabilities include:
For more information, see CodeMender overview.
Open model endpoint deprecations
The following open model endpoints are deprecated and will be retired on October 21, 2026. For more information, see Open model deprecations.
deepseek-ocr-maasdeepseek-r1-0528-maasdeepseek-v3.2-maasdeepseek-v3.1-maasglm-5-maasglm-4.7-maasgpt-oss-20b-maaskimi-k2-thinking-maasllama-3.3-70b-instruct-maasminimax-m2-maasmultilingual-e5-large-instruct-maasmultilingual-e5-small-maasqwen3-235b-a22b-instruct-2507-maasqwen3-coder-480b-a35b-instruct-maasqwen3-next-80b-a3b-instruct-maasqwen3-next-80b-a3b-thinking-maasCodeMender is available in Public Preview
CodeMender is now available in Public Preview for select customers. CodeMender is an AI-powered code security agent that runs on your local workstation to find, verify, and fix security vulnerabilities in your codebase.
Key capabilities in this release include:
For more information, see the CodeMender overview.
As of July 13, 2026, Looker reports have been deprecated. If you had previously enabled the preview for Looker reports, be aware of the following:
Access to the rest of your Looker content in your instance will remain unaffected and you will continue to have access to Looker as a data source from Data Studio and Data Studio Pro.
You can create ad hoc Explores using Looker's Self-service Explores feature, which lets you upload CSV, XLS, and XLSX files to Looker and then query and visualize the data in a Looker Explore without needing to configure a LookML model or set up Git version control.
For Google Cloud resources that are registered as App Hub workloads or services, VPC Flow Logs records contain application-specific labels. For more information, see App Hub labels.
]]>Starting on the following dates, you can no longer create a job that locates its Compute Engine resources outside of the job's location.
allowedLocations[] field with any region or zones
outside of the job's location, changes are starting on June 30, 2027.If none of your jobs specify the allowedLocations[] field, then no action is
required. Otherwise, ensure that any region or zones specified in the
allowedLocations[] field are in the same region as the job's location
before these dates. For more information, see
Batch locations.
Lakehouse for Apache Iceberg: Data Products with special characters, such as "/" or "-", are not supported and will not be available in BigQuery even if shared from SAP BDC to BigQuery. If you share a Data Product with special characters, this could cause the refresh to stop and require re-enrollment. Known SAP systems producing these Data Products include SAP Business Warehouse (BW) sources and SAP SuccessFactors.
Lakehouse for Apache Iceberg: Cross-cloud Lakehouse now supports integration with SAP Business Data Cloud (BDC) in Preview.
This update includes the following features:
For more information, see Set up cross-cloud Lakehouse for SAP BDC, Query SAP BDC data, and Publish Data Products to SAP BDC.
For regional external passthrough Network Load Balancers, you can reserve specific or automatically allocated bring your own IP (BYOIP) IPv6 addresses before creating a load balancer, so that the IPv6 address persists independently of the load balancer's lifecycle. You can also promote an ephemeral BYOIP IPv6 address that is in use by a load balancer to a reserved static IP address.
For more information, see the following documentation:
This feature is in Preview.
You can now use the WildFire service to protect your network against unknown, novel malware, and file-based threats. WildFire integrates advanced malware sandboxing and real-time machine learning (ML) to perform deep inspection of network-routed file transfers and block zero-day malware before it reaches your workloads. WildFire is available in the Cloud Firewall Enterprise tier.
For more information, see WildFire overview and Configure WildFire in your network. This feature is available in Preview.
Cloud SQL for PostgreSQL now supports authentication via Secret Manager when executing SQL statements using the Data API (executeSql). You can store your database password in a regional secret in Secret Manager and pass the secret version resource name in your API request.
For more information, see Execute SQL statements on a Cloud SQL instance.
Object Lifecycle Management conditions for
sizeAboveBytes and sizeBelowBytes
let you define a minimum and maximum size threshold for lifecycle actions.
Encrypting disks, snapshots, images, and machine images with customer-supplied encryption keys (CSEKs) is deprecated and will be disabled on July 20, 2027.
For more information and alternatives to CSEKs for your Compute Engine resources, see Deprecation of customer-supplied encryption keys (CSEK) in Compute Engine.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.6.143 | v24.0.9 | v1.7.34 | See List |
Fixed a bug in the XFS file system where direct I/O writes could use outdated block mappings during Copy-on-Write operations.
Upgraded net-misc/curl to 8.21.0.
Fixed CVE-2026-58013,CVE-2026-58014,CVE-2026-58015,CVE-2026-58016 in glib.
Fixed CVE-2025-13462 in dev-lang/python
Fixed CVE-2026-3644 in dev-lang/python
Fixed CVE-2026-4224 in dev-lang/python
Fixed CVE-2026-43010 in the Linux kernel.
Fixed CVE-2026-46135 in the Linux kernel.
Fixed CVE-2026-46331 in the Linux kernel.
Fixed CVE-2026-53163 in the Linux kernel.
Fixed CVE-2026-53167 in the Linux kernel.
Fixed CVE-2026-58469, CVE-2026-58471, CVE-2026-58472 in net-misc/wget.
Updated containerd to v1.7.34. This resolves CVE-2026-46680 and CVE-2026-53488.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.6.143 | v27.5.1 | v2.0.10 | See List |
Updated containerd to v2.0.10. This resolves CVE-2026-46680.
Fixed a bug in the XFS file system where direct I/O writes could use outdated block mappings during Copy-on-Write operations.
Upgraded net-misc/curl to 8.21.0.
Fixed CVE-2026-58013,CVE-2026-58014,CVE-2026-58015,CVE-2026-58016 in glib.
Fixed CVE-2026-23278 in the Linux kernel.
Fixed CVE-2026-43010 in the Linux kernel.
Fixed CVE-2026-46135 in the Linux kernel.
Fixed CVE-2026-46331 in the Linux kernel.
Fixed CVE-2026-53163 in the Linux kernel.
Fixed CVE-2026-53167 in the Linux kernel.
Fixed CVE-2026-58469, CVE-2026-58471, CVE-2026-58472 in net-misc/wget.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.12.94 | v27.5.1 | v2.2.5 | See List |
Fixed a bug in the XFS file system where direct I/O writes could use outdated block mappings during Copy-on-Write operations.
Upgraded net-misc/curl to 8.21.0.
Fixed CVE-2026-58013,CVE-2026-58014,CVE-2026-58015,CVE-2026-58016 in glib.
Fixed CVE-2026-43216 in the Linux kernel.
Fixed CVE-2026-58469, CVE-2026-58471, CVE-2026-58472 in net-misc/wget.
Updated containerd to v2.2.5. This resolves CVE-2026-46680,CVE-2026-50195,CVE-2026-53492,CVE-2026-53488.
| Kernel | Docker | Containerd | GPU Drivers |
| COS-6.12.94 | v27.5.1 | v2.1.9 | See List |
Upgraded net-misc/curl to 8.21.0.
Fixed CVE-2026-58013,CVE-2026-58014,CVE-2026-58015,CVE-2026-58016 in glib.
Fixed CVE-2026-43010 in the Linux kernel.
Fixed CVE-2026-43216 in the Linux kernel.
Fixed CVE-2026-58469, CVE-2026-58471, CVE-2026-58472 in net-misc/wget.
Updated containerd to v2.1.9. This resolves CVE-2026-46680,CVE-2026-50195, CVE-2026-53492,CVE-2026-53488.
You can now view, manage, and deploy Firestore Security Rules directly in the Google Cloud console for Firestore (Standard and Enterprise editions). You can create new rulesets and clone or restore rulesets from the timeline.
To learn more, see Manage Firestore Security Rules or Use the Google Cloud console.
Various bug fixes and minor product enhancements.
Security update for Server-Side Request Forgery (SSRF) in Agent Studio
This release fixes a Server-Side Request Forgery (SSRF) vulnerability in the
auto-generated /api-proxy backend endpoint for web applications created
before July 1, 2026, using Agent Studio.
If you downloaded, generated, or deployed web application code from Agent Studio before July 1, 2026, regenerate the app from Agent Studio and deploy the new version. For more information, see Quickstart: Deploy your Agent Studio prompt as a web application
The updated backend code includes strict domain allowlist validation, ensuring
that destination hostnames for the /api-proxy endpoint end with allowed Google
Cloud domains, such as *-aiplatform.clients6.google.com
Google Distributed Cloud (software only) for bare metal 1.35.300-gke.87 is now available for download. To upgrade, see Upgrade clusters. Google Distributed Cloud for bare metal 1.35.300-gke.87 runs on Kubernetes v1.35.3-gke.400.
After a release, it takes approximately 7 to 14 days for the version to become available for installations or upgrades with the GKE On-Prem API clients: the Google Cloud console, the gcloud CLI, and Terraform.
If you use a third-party storage vendor, check the Google Distributed Cloud-ready storage partners document to make sure the storage vendor has already passed the qualification for this release of Google Distributed Cloud for bare metal.
The following change was added in 1.35.300-gke.87:
csi-snapshot-validation-webhook component. Upstream Kubernetes validation is now handled natively via Common Expression Language (CEL) rules within the deployed Custom Resource Definitions (CRDs). For more information, see Volume snapshots.
The following issues were fixed in 1.35.300-gke.87:
To improve security, Ubuntu node images in GKE version 1.37 and later don't
pre-install the vulkan-tools package. If you run Vulkan diagnostic tools
(such as vulkaninfo) directly on GKE Ubuntu hosts, then you must manually
install the vulkan-tools package. This change doesn't affect containerized
GPU/Vulkan workloads.
[Spotlight Feature] Deprecation of Google Security Operations legacy SIEM APIs
Google Security Operations is deprecating its legacy SIEM APIs—Backstory API (including Customer Management API) and Ingestion API—in favor of the modern Chronicle API.
Key dates
This change applies only to custom scripts, integrations, SOAR connectors, or ingestion feeds calling legacy Backstory API or Ingestion API endpoints. Any changes impacting the Google SecOps UI are already addressed and don't call for your action.
Next steps
Audit API usage to identify any affected components that currently call legacy Backstory API or Ingestion API endpoints, and replace them with Chronicle API endpoints.
Validate and test that your updated components work properly.
For more information, see Migrate from legacy API to Chronicle API.
[Spotlight Feature] Deprecation of Google Security Operations legacy SIEM APIs
Google Security Operations is deprecating its legacy SIEM APIs—Backstory API (including Customer Management API) and Ingestion API—in favor of the modern Chronicle API.
Key dates
This change applies only to custom scripts, integrations, SOAR connectors, or ingestion feeds calling legacy Backstory API or Ingestion API endpoints. Any changes impacting the Google SecOps UI are already addressed and don't call for your action.
Next steps
Audit API usage to identify any affected components that currently call legacy Backstory API or Ingestion API endpoints, and replace them with Chronicle API endpoints.
Validate and test that your updated components work properly.
For more information, see Migrate from legacy API to Chronicle API.
Preview: You can reserve static external IPv6 addresses from
bring your own IP addresses (BYOIP) sub-prefixes that are in
EXTERNAL_IPV6_FORWARDING_RULE_CREATION mode.
You can assign these addresses to forwarding rules for external passthrough Network Load Balancers and external protocol forwarding. You can also promote ephemeral IPv6 BYOIP addresses that are used by external forwarding rules to reserved static IP addresses.
For more information, see Create external forwarding rules.
]]>Release 6.3.94 is being rolled out to the first phase of regions as listed here.
This release contains internal and customer bug fixes.
]]>Release 6.3.93 is now available for all regions.
]]>