You can see the latest product updates for all of Google Cloud on the Google Cloud page, browse and filter all release notes in the Google Cloud console, or programmatically access release notes in BigQuery.
To get the latest product updates delivered to you, add the URL of this page to your feed reader, or add the feed URL directly.
November 05, 2026
The AlloyDB Omni Kubernetes operator version 1.6.4 is generally available (GA) and adds support for AlloyDB Omni version 15.18.1 for PostgreSQL 15 clusters.
To get all the bug fixes and updates in this release, use operator 1.6.4 with database version 15.18.1.
For information about how to upgrade an existing PostgreSQL 15 cluster to database version 15.18.1, see Upgrade the AlloyDB Omni Kubernetes operator and Perform a minor database version upgrade.
Fixed an issue where the health check prober might report a false failure status because it kept using cached client connections after an operator client certificate expired or was revoked. For more information, see Mitigate operator client certificate expiration in operator versions earlier than 1.8.2.
Fixed an issue where audit logger pipe file descriptors might leak when the
database restarted with pgaudit enabled. The leak left orphaned logger
processes that might cause out-of-memory (OOM) conditions.
This fix is available only for PostgreSQL 15 clusters running database version 15.18.1 on operator 1.6.4. For more information, see Perform a minor database version upgrade.
Updated the tzdata version 2026e time zone data in AlloyDB Omni version
15.18.1 to incorporate the 2026 IANA daylight saving time rules.
This update is available only for PostgreSQL 15 clusters running database version 15.18.1 on operator 1.6.4. If your database uses an affected time zone, see Update timestamps and indexes after you upgrade to 15.18.1.
Removed the distrusted emSign ECC Root CA - G3 and AAA Certificate Services
root certificate authority (CA) certificates from the operating system trust
store in AlloyDB Omni version 15.18.1. The trust store now uses
ca-certificates version 20250419~deb12u2. For details, see security
bulletin GCP-2026-071.
This update is available only for PostgreSQL 15 clusters running database version 15.18.1 on operator 1.6.4. For more information, see Perform a minor database version upgrade.
October 30, 2026
In AlloyDB Omni Kubernetes operator versions earlier than 1.8.2, the operator
doesn't reload its renewed dbs-al-cert-l2-operator mutual TLS (mTLS) client
certificate into existing high availability (HA) health-check connections.
Approximately 60 days after the certificate is issued, health checks might fail
with a tls: expired certificate error and trigger an unplanned failover. To
prevent this issue, upgrade to operator version 1.8.2 or later. Until you
upgrade, rotate the dbs-al-cert-l2-operator Secret in the
alloydb-omni-system namespace and restart the local-controller-manager
Deployment at least 10 days before the certificate expires. For instructions,
see
Mitigate operator client certificate expiration in operator versions earlier than 1.8.2.
October 21, 2026
The AlloyDB Omni web user interface is generally available (GA) for AlloyDB Omni version 18.1.0 and later. The web UI provides a centralized graphical dashboard to monitor, analyze, and manage your AlloyDB Omni deployments, including the following capabilities:
- Connect to the web UI: set up multiple connection profiles to connect to multiple clusters. For more information, see Install the AlloyDB Omni web user interface and Connect to the AlloyDB Omni web user interface.
- Monitor cluster high availability: view cluster topology, instance roles, node health, and replication status. For more information, see Monitor cluster high availability.
- Monitor system resources with system insights: track real-time and historical CPU, memory, storage, and disk I/O metrics. For more information, see View system metrics using system insights.
- Analyze query performance with query insights: identify slow queries and high disk I/O statements, and inspect SQL execution statistics. For more information, see View query performance using query insights.
- Monitor the columnar engine: view the column store inventory, analyze columnar query execution statistics, and track column store memory usage. For more information, see Monitor the columnar engine.
- Audit database parameters: search, filter, and audit active database parameters and identify customized flags that differ from default engine values. For more information, see Audit database parameters using the AlloyDB Omni web user interface.
- Monitor AI and ML capabilities: view registered machine learning models and external model endpoints available to the database engine; check connectivity status with remote AI endpoints, such as Vertex AI, and local model inference runtimes; and view the status of vector search indexing, vector dimensionality, and index build health.
October 13, 2026
The AlloyDB Omni Kubernetes operator version 1.9.0 is generally available (GA). Version 1.9.0 includes the following features and changes:
AlloyDB Omni provides support for the following new PostgreSQL minor versions:
- The
DBClustercustom resource now supports configuring CPU and memory resource limits for PgBouncer and system control plane agents such as thelogrotateagent. - The
UserDefinedAuthenticationresource now supports configuring custompg_hba.confentries. This lets you customize client authentication behavior, such as restricting client subnets, specifying custom authentication methods, or configuring user mapping. - The
DBClustercustom resource now supports configuring an image pull secret forDBClusterservice accounts. You can operate with private image registries to automatically pull pod images without needing to manually pre-create service accounts. For more information, see Install AlloyDB Omni using the container orchestrator. - The
TdeConfigcustom resource now supports Thales CipherTrust Manager as an external Key Management Service (KMS) provider for Transparent Data Encryption (TDE), in addition to HashiCorp Vault. For more information, see Create a TDE-enabled cluster. - Cross-data-center replication now supports auto-heal of standbys. For more information, see Work with cross-data-center replication.
- The
DBClustercustom resource now supports using external, non-cert-manager custom issuer kinds, for example,VenafiIssuerfrom thejetstack.ioAPI group. For more information, see Configure custom issuers.
Resolved the following security vulnerabilities in the operator images:
- CVE-2026-54370
- CVE-2026-54369
- CVE-2026-6238
- CVE-2026-5928
- CVE-2026-5450
- CVE-2026-5435
- CVE-2026-4438
- CVE-2026-4437
- CVE-2026-4046
Backported fixes for the following vulnerabilities in the kernel images:
Resolved an issue where the logrotate-agent could fail or hang during high audit log volumes, such as when pgaudit.log is set to all, which could lead to logrotate-agent out-of-memory errors, unbounded log growth, and disk exhaustion. The log rotate workflow has been improved, and resource limits for sidecar containers are now configurable to prevent resource constraints. For more information about the operator's logging mechanism and supported parameters, see Configure log rotation.
After you enable cross-data-center replication, the monitoring agent might encounter permission errors when attempting to gather cgroups information. If this error occurs, delete the downstream database Pod so that the operator recreates and restarts the Pod.
The AlloyDB Omni Kubernetes operator uses cert-manager, a third-party open source component, to provision Transport Layer Security (TLS) certificates. In cert-manager version 1.21.2 and earlier, a known race condition might cause certificate issuance to fail. For more information, see Resolve certificate issuance failures in cert-manager.
October 05, 2026
The AlloyDB Omni Kubernetes operator version 1.8.2 is now available.
This release includes security fixes for the following vulnerabilities:
Fixed an issue in the Node Manager where the -nm pods failed to schedule due to those pods not inheriting scheduling config from the instance. Node Manager is a component that lets you roll out control plane features without requiring a data plane restart or database downtime.
Fixed a concurrent read-modify-write race condition on pg_hba.conf between control plane and data plane agents.
Fixed an issue where the health check prober could post a false failure status because it did not hot-reload cached client connections when a certificate expired or was revoked. If you run an operator version earlier than 1.8.2, see Mitigate operator client certificate expiration in operator versions earlier than 1.8.2.
Fixed an issue where the database failed to start due to a multiple primary key constraint for g_columnar_config in the post-startup script.
June 18, 2026
AlloyDB Omni documentation now includes documentation for the following AlloyDB AI features available in GA:
Auto vector embeddings automate the lifecycle of vector embeddings for large datasets, keeping them in sync with transactional data using real-time or batch updates. This feature includes a bulk mode that performs refreshes and migrations up to 130x faster than row-by-row processing."
AI functions integrate LLMs like Gemini to bring world knowledge to your AlloyDB data and incorporate advanced semantic search and ranking capabilities directly into your SQL workflows.
Use array-based AI functions to improve the efficiency of large-scale AI operations by enabling batch processing of natural language prompts in your SQL queries. For more information, see Perform intelligent SQL queries using AI functions.
AlloyDB Omni documentation now includes documentation for the following AlloyDB AI features, which are available in Preview:
- The AlloyDB columnar engine provides a read-optimized, in-memory cache for HNSW indexes, increasing the number of queries per second (QPS) that your database can handle for vector search workloads. For more information, see Accelerate vector search with the columnar engine.
- Hybrid search supports Elasticsearch, which lets you use the
external_search_fdwextension to connect to Elasticsearch and perform hybrid searches in AlloyDB. For more information, see Access Elasticsearch data from AlloyDB. - The sentiment analysis and summarization functions let you process and analyze unstructured data in your database. For more information, see Analyze text sentiment and Summarize content.
- You can defer ScaNN index creation on an empty table or a table with insufficient rows until the table has sufficient data. For more information, see Create a ScaNN index.
- The alloydb_scann extension now supports four-level tree indexes, providing support for tables with up to 10 billion vector rows. For more information, see Four-level ScaNN tree indexes.
June 03, 2026
The AlloyDB Omni Kubernetes operator version 1.7.1 is generally available (GA) and includes the following vulnerability fixes:
- CVE-2026-32288 by upgrading to Golang Version 1.26.2.
- CVE-2026-4878 (
libcap). - CVE-2026-40356 (
krb5-libs) in theg-pgbouncercontainer image.
May 08, 2026
The AlloyDB Omni Kubernetes operator (version 1.7.0) is officially certified for Red Hat OpenShift. The operator and its associated database images are validated to run optimally and securely on Red Hat OpenShift.
This release includes the following:
- OpenShift Certified Catalog: the AlloyDB Omni Operator is now officially listed on the OpenShift Certified Catalog.
- Built-in management: discover, install, and manage the operator using OpenShift's built-in Operator Lifecycle Manager (OLM) user interface.
- OLM support: support for OLM bundles is available, facilitating a seamless, one-click installation experience and automated management for enterprise deployments. For more information, see Install the AlloyDB Omni operator.
- This certification is supported by AlloyDB Omni versions starting with 16.11.0, 17.7.0, and 18.1.0.
April 22, 2026
The naming conventions for AlloyDB Omni deployment options are updated to more
accurately reflect their respective installation and orchestration methods.
The following table summarizes the primary name changes:
| Original name | New name |
|---|---|
| AlloyDB Omni for containers | AlloyDB Omni using containers |
| AlloyDB Omni for Kubernetes | AlloyDB Omni using the container orchestrator |
| AlloyDB Omni for Linux | AlloyDB Omni using RPM |
| Red Hat RPM orchestrator | AlloyDB Omni using the RPM orchestrator |
This update clarifies the distinction between AlloyDB Omni deployment environments (such as Kubernetes or standalone Linux) and the packaging and orchestration technologies used to manage them. For more information, see Choose your deployment environment.
April 09, 2026
The AlloyDB Omni Kubernetes operator version 1.7.0 is generally available (GA) and includes the following features and changes:
- AlloyDB Omni supports PostgreSQL 18.1.0. For more information, see AlloyDB Omni Kubernetes operator. This operator version also supports AlloyDB Omni minor versions 15.15.0, 16.11.0, and 17.7.0.
- Enterprise readiness for AlloyDB Omni on Google Distributed Cloud Software Only lets the operator manage the full database lifecycle, including provisioning, high availability, and disaster recovery, directly on your own physical bare metal infrastructure.
- Database pods feature a hardened security posture with restricted
Kubernetes
securityContextsettings enabled by default, including non-root execution and read-only root filesystems. - Added support for Transparent Data Encryption (TDE) (Preview), providing at-rest encryption for database clusters. This is available beginning with AlloyDB Omni 18.1.0. For more information, see Transparent Data Encryption.
- Added support for custom configuration of database port numbers for your clusters. For more information, see Configure custom ports.
- Introduced a framework for defining and collecting custom database metrics to enhance monitoring capabilities. For more information, see Configure custom metrics.
- Added detailed memory allocation metrics for improved memory and performance monitoring. For more information, see AlloyDB Omni metrics.
- Improved Active Directory integration
with support for
hostsslconnection types, multi-domain LDAP base DNs, and customizable LDAP search attributes. For more information, see Integrate Active Directory. - Added support for rewind trigger functionality in symmetric disaster recovery configurations. For more information, see Work with cross-data-center replication.
- You can perform CPU and memory scaling operations with minimal downtime using Low Downtime Maintenance (Preview). For more information, see Scale a database cluster.
- Additional database flags are supported.
- All current and earlier versions of Helm charts are available in the Artifact Registry. However, Helm charts are no longer available in Cloud Storage.
- Updated manual major version upgrade instructions to use versioned data directory paths. For more information, see Upgrade database major version.
- Updated Choose compatible Kubernetes operator and database cluster versions with latest versions.
The AlloyDB Omni Kubernetes operator now automatically scales the wal_keep_size and max_wal_size database flags based on your configured storage size to optimize storage and checkpointing behavior. If you manually configure these flags, your custom values are respected and override the system-selected defaults.
Reliability improvement for backups from standby instances: Fixed an issue in the AlloyDB Omni Kubernetes operator version 1.7.0 where backups taken from a standby instance could be incomplete if the primary instance were under heavy load. The backup process now performs a checkpoint before starting to ensure data integrity.
Upgrade required: If you use the backup from standby feature, upgrade to the latest versions of the AlloyDB Omni control plane and data plane. After you perform the upgrade, take a new backup to make sure that your data is fully protected.
Restore operations might fail if a database backup is running at the same time. If this occurs, retry the restore operation after some delay.
September 18, 2024
The AlloyDB Omni operator is now available in Preview on Google Distributed Cloud (GDC) connected. For more information, see Install AlloyDB Omni on Kubernetes.
September 16, 2024
The postgres_ann extension has been renamed to alloydb_scann. Before you upgrade to AlloyDB Omni Kubernetes operator version 1.1.1, you must drop any indexes created using the earlier postgres_ann version, then upgrade AlloyDB Omni, and then create the indexes again using the alloydb_scann extension.
AlloyDB Omni Kubernetes operator version 1.1.1 is now available. This patch fixes the following issues:
- Fixed a regression for the AlloyDB Vertex AI integration.
- Fixed a bug in which upgrading from version 1.0.0 to version 1.1.0 failed when using injected sidecars.
- Fixed a bug in which backups weren't reestablished correctly across failovers when using the Commvault sidecar with high availability (HA) configurations.
- Fixed a bug that caused a status to be incorrectly set by the load balancer, resulting in erroneous reports that the database cluster wasn't ready.
Added a tutorial that shows you how to set up a connection from an application running in a Google Kubernetes Engine autopilot cluster to an AlloyDB instance.
Upgrading to version 1.1.1 of the AlloyDB Omni Kubernetes operator might result in a brief interruption to all database clusters. No data loss is expected.
August 22, 2024
AlloyDB Omni now offers in-depth documentation that describes how to install and use AlloyDB Omni in virtual machine (VM) environments. The information in these documents is in addition to the AlloyDB Omni documentation that describes how to install and use AlloyDB Omni in VM and Kubernetes environments. For more information, see Get started with AlloyDB Omni on VMs.
The extension temporal_tables version 1.2.2 has been added to extensions supported by AlloyDB.
The following extensions are updated:
- Updated
pg_partmanto version 4.7.4. - Updated
pgttto version 3.0.0.
August 16, 2024
August 05, 2024
AlloyDB Omni version 15.5.5 is now generally available (GA). This version includes the following features and changes:
The AlloyDB Omni Kubernetes Operator version 1.1.0 is generally available (GA) and it includes the following new features:
- Cross-data center replication to support automated setup of disaster recovery.
- Support for the Red Hat OpenShift container platform version 4.14 and later in Preview.
- Variable number of nodes per read pool instance.
- Customizable count for the number of failed checks before auto-failover kicks in for a High Availability (HA) configuration.
- Version 1.1.0 of the Operator is supported on AlloyDB Omni version 15.5.5 and later.
- A dedicated backup server for streamlined backup operations.
- When you create a database cluster, the admin password secret is no longer deleted after you specify it.
The
postgres_annextension is renamed toalloydb_scann. Before you upgrade AlloyDB Omni, you must drop any indexes created using the olderpostgres_annversion, then upgrade AlloyDB Omni, and then create the indexes again using thealloydb_scannextension.Various bug fixes.