Use release channels for Google Kubernetes Engine (GKE) to pick versions for your clusters with your chosen balance between feature availability and stability.
GKE automatically upgrades all clusters over time, including those not enrolled in a release channel, to ensure that they receive security updates, fixes to known issues, new features, and run a supported Kubernetes version. You can control the timing of upgrades with maintenance windows and exclusions.
We recommend enrolling your cluster in a release channel, as this gives you the most control with regards to the scope of maintenance exclusions—preventing specific types of upgrades instead of all upgrades—and sequencing the rollout of cluster upgrades. Autopilot clusters can only be enrolled in a release channel.
If you don't enroll your Standard cluster in a release channel, you can disable node auto-upgrades for selected node pools and manually manage upgrades to the nodes in these node pools. However, all clusters' control planes are automatically upgraded, and when a version reaches the end of support, cluster control planes and nodes are automatically upgraded regardless of release channel enrollment. To learn more, see the comparison between clusters enrolled and not enrolled in a release channel.
What channels are available
The following table explains the properties of available release channels, each offering a trade-off between feature availability and stability. All channels offer supported releases of GKE and are considered generally available (GA), although individual features might not always be GA, as marked. The Kubernetes releases in these channels are official Kubernetes releases and include both GA and beta Kubernetes APIs.
| Channel | New Kubernetes release availability | When to use this channel |
|---|---|---|
| Rapid | Several weeks after upstream open source GA | Get the latest Kubernetes release as early as possible, and be able to use new GKE features the moment they go GA. GKE upgrades your cluster frequently to stay on the latest available patch version, and deliver newer Kubernetes capabilities. Clusters subscribed to the Rapid channel use GA versions, however we recommend to use the Rapid channel to test newer Kubernetes versions and API on pre-production environments. |
| Regular (default) | 2-3 months after releasing in Rapid | Access GKE and Kubernetes features soon after they go GA, but on a version that has been qualified over a longer period of time. Offers a balance of feature availability and release stability, and is what we recommend for most users. |
| Stable | 2-3 months after releasing in Regular | Prioritize stability over new features. GKE rolls out changes and new versions in this channel last, after being validated on the Rapid and Regular channels, which allows even more time for validation. | Extended | Aligned with Regular channel | Use this channel to get long-term support, keeping a cluster running a minor version for as long as possible. To learn more, see Get long-term support with the Extended channel. |
| No channel (not recommended) | Aligned with Regular channel | This option is not recommended. Control plane and nodes are auto-upgraded in alignment with the Regular and Stable channels. When you need to disable node auto-upgrades at a node pool level, you can use this configuration option. Alternatively, with release channels you can disable node auto-upgrades at a cluster level. For more details, see the comparison between clusters enrolled and not enrolled in a release channel. |
When you enroll a cluster in a release channel, that cluster is upgraded automatically on or after the date specified in the Auto Upgrade column of the GKE release schedule.
When a version has accumulated usage and demonstrated stability across clusters in the Rapid channel, it is promoted to the Regular channel. Eventually, the version is promoted to the Stable channel, which only receives high-priority updates. Each promotion signals a graduating level of stability and production-readiness, based on observed performance of clusters running that version.
Critical security patches are delivered to all release channels, to protect your clusters and Google's infrastructure. In rare emergency situations, GKE might auto-upgrade clusters to versions that aren't yet available in their release channel.
What versions are available in a channel
Each release channel offers multiple minor versions. These versions have met the qualification standards for that specific channel. This set of available versions includes a default version for cluster creation, which is selected from a set of available versions from that channel. To see what versions are available in a release channel, follow the instructions to view the default and available versions for release channels.
- New patch releases become available at least one week prior to becoming an auto-upgrade target for all channels.
- New minor releases become available:
- at least two weeks prior to becoming an auto-upgrade target for the Rapid channel.
- at least four weeks prior to becoming an auto-upgrade target for the Regular and Stable channels.
You can test a newly available GKE version prior to upgrading your production environment. For example, you can subscribe to upgrade notifications to be informed of newly available versions, and then proactively upgrade a pre-production environment to the new version before it becomes the auto-upgrade target for clusters in your production environment.
If you need to keep a cluster on a specific version, for example to validate or test newer versions prior to upgrading, we recommend using maintenance exclusions.
After a minor version has been made available in a release channel, it will remain available in that release channel for new or existing clusters until it reaches its end of support date.
What happens when a version becomes an auto-upgrade target in a release channel
The GKE release notes lists new auto-upgrade targets for clusters that run specific minor versions with Version updates such as the 2024-R33 note. When new auto-upgrade targets are available, GKE assesses if your specific cluster should be upgraded to that new version. GKE considers if your cluster has maintenance policies or other constraints preventing the automatic upgrade, and won't ignore those reasons except in rare emergency situations. To get auto-upgrade targets for a specific cluster, see Get information about a cluster's upgrades (Preview).
If 10 days have passed after a new version became an auto-upgrade target for your cluster's minor version in the release channel, and automatic upgrades haven't started for your cluster, the delay might be due to one of the following reasons:
- Your cluster is temporarily ineligible for automatic upgrades. This might
occur because:
- The cluster is outside the timeframe of a configured maintenance window.
- The cluster is within the timeframe of a maintenance exclusion.
- Automatic upgrades are paused because your cluster uses deprecated Kubernetes features that are removed in the next minor version.
- The cluster was automatically upgraded to a patch version less than 24 hours ago.
- The cluster was automatically upgraded to a minor version less than 30 days ago and the auto-upgrade target is a new minor version.
- GKE paused the rollout of the new auto-upgrade target for
technical or business reasons:
- Technical issues were discovered with the new version.
- A production freeze is in place due to a critical business season, such as Black Friday.
What is the default version?
When you create a cluster in a release channel, by default, the cluster uses the default patch version for cluster creation in the selected release channel. However, you can specify a different available version when you create a cluster in the release channel.
Release channels have multiple minor versions available in a release channel, and the default version can sometimes be one of the auto-upgrade targets for a release channel.
Run patch versions from a newer channel
In addition to the listed available patch versions for a release channel, you can run patch versions from newer release channels than the one where your cluster is enrolled if the minor version is available in the cluster's release channel and you use the gcloud CLI or the GKE API.
For example, if the following versions were available in the Rapid and Regular channels:
- Rapid: 1.23.2-gke.700, 1.22.4-gke.1500
- Regular: 1.21.4-gke.400, 1.22.1-gke.400
A cluster enrolled in the Regular channel that is running GKE version 1.22.1-gke.400 could upgrade to 1.22.4-gke.1500, but not to 1.23.2-gke.700 as it is a different minor version.
To upgrade to a patch version on a newer channel, your cluster's control plane must run a patch release with the same minor version. For example, if the cluster is running 1.21.3-gke.200, you must first upgrade the cluster to a patch version available in its current release channel, 1.22.1-gke.400. You can then upgrade the cluster to 1.22.4-gke.1500.
You can also create a new cluster that runs 1.22.4-gke.1500 and is enrolled in the Regular channel.
GKE keeps the cluster on the patch version from the newer channel until a newer auto-upgrade target becomes available in the enrolled channel for the cluster.
Learn what's new in a channel
To learn what's new in a release channel, review the release notes. There are separate release notes for each release channel, in addition to the overall release notes.
| Release channel | Release notes |
|---|---|
| Rapid channel | HTML or Atom feed. |
| Regular channel | HTML or Atom feed. |
| Stable channel | HTML or Atom feed. |
Choose the best release channel for your cluster
Channels include only GA versions of Kubernetes, and each channel represents a different level of quality and maturity of the Kubernetes and GKE releases. The following diagram illustrates the adoption cycle for release channels:
As this diagram shows, the available release channels use versions in the middle of the adoption cycle, including Early adopters (Rapid channel), Early majority (Regular channel), and Majority (Stable channel). The earliest part of the adoption cycle is the Innovators, who test the newest features using the upstream Kubernetes release. The last part of the adoption cycle is the Late majority, where you are using a version that is nearing deprecation and need to transition to a supported version.
In your pre-production environments, use the Rapid channel for newer versions where you can test features as soon as they are generally available.
For production workloads that require maturity over newer features, we recommend using the Regular (default) channel or the Stable channel.
- If you need to closely track new features, consider using the Regular channel, which offers a balance between stability and freshness of the Kubernetes OSS version.
- If your requirement is maturity, especially for production clusters, use the Stable channel.
GKE upgrades clusters to a newer version that meets the channel's quality standards. However, we recommend upgrading your clusters ahead of time because this provides you with:
- Better control of your upgrades, and alignment with your working hours.
- Better predictability because GKE skips auto-upgrading clusters that meet the release target (i.e., clusters that were manually upgraded to the next target version). Nodes are automatically upgraded to the recommended version in their selected channel to align with the control plane version and to protect you from vulnerabilities and unsupported version skew.
Get long-term support with the Extended channel
GKE provides long-term support for Kubernetes minor versions through the Extended channel. With this channel, you can stay on a minor version for up to 24 months. After the 14-month standard support period, your cluster receives security patches for approximately 10 more months during the extended support period.
How GKE automatically upgrades clusters in the Extended channel
For clusters enrolled in the Extended channel, GKE automatically upgrades clusters in the following way:
- During the standard support period: GKE upgrades clusters to newer patch versions of the same minor version following the same cadence as the Regular channel.
- During the extended support period: GKE continues to provide security patch updates for the minor version. Around 2 months before the end of extended support, GKE begins upgrading clusters to the next minor version, unless the cluster is using deprecated features or APIs. You can use maintenance exclusions to prevent minor version upgrades until the end of extended support.
- At the end of extended support: GKE upgrades all clusters still running the now-unsupported minor version, regardless of blocking issues. You can't configure maintenance exclusions past this date.
To learn more about the different periods of availability and what upgrades GKE provides during the extended support period, see GKE minor version lifecycle.
Limitations for enrolling a cluster in the Extended channel
You can only enroll Standard clusters running version 1.27 or later in the Extended channel. Both the cluster control plane and nodes must run 1.27 or later.
You can't enroll a cluster in the Extended channel when using the following features:
- Autopilot cluster mode
- Alpha clusters
- Explicitly-enabled Kubernetes beta APIs
- Gateway
- Windows Server node pools
- Customized sysctl configuration options
- Backup for GKE
- Config Connector
- The following multi-cluster features:
Pricing for extended support
If you want to enroll a cluster in the Extended channel, ensure that you've reviewed the pricing for extended support. You can enroll a cluster in the Extended channel for no additional cost if the project has enabled GKE Enterprise. Or, for GKE Standard edition clusters, pay-per-use costs apply when your cluster is enrolled in the Extended channel and your cluster's minor version enters the extended support period.
Best practices for the Extended channel
Review the following scenarios to understand how you can use the Extended channel to minimize disruption from minor version upgrades.
The supported scenarios require some manual action over time to get the most benefit from the channel. We don't recommend that you enroll a cluster in the channel without plans to move to the next minor version, as GKE will eventually upgrade the cluster to newer minor versions with the same frequency as other channels, and your cluster will potentially incur greater cost and receive new features last.
Supported and unsupported scenarios
To learn more about supported and unsupported scenarios, review the following table, and see Use the Extended channel when you need long-term support for more details. A checkmark () indicates a supported scenario:
| Upgrade scenario | Supported | Summary | Time between minor version changes | Manual action required |
|---|---|---|---|---|
| Temporarily stay on minor version for longer | Stay on a minor version to mitigate an issue preventing upgrades. | Same average frequency, with interruption for extra time on one minor version. |
|
|
| Manually upgrade minor version 1-2 times a year | Get new features but with less frequent minor upgrades, when it's optimal for the workloads in the cluster. | 1-2 times per year. |
|
|
| Do nothing and receive minor upgrades with same frequency | Get minor version upgrades with the same frequency as other channels, and new features as late as possible. | Every 4 months, on average. |
|
Clusters not enrolled in a release channel
We don't recommend this configuration option due to the limitations with clusters not enrolled in release channels, but you can choose not to enroll a Standard cluster in a release channel (known as "no channel" and "static"). Don't use this option unless individual node pools can't be automatically upgraded, and you must manually upgrade those nodes instead. If your cluster is not enrolled in a release channel, you can disable node auto-upgrade for selected node pools. With release channels, you can achieve the same outcome at the cluster level across all node pools. However, regardless of release channel enrollment, all clusters' control planes are automatically upgraded, and when a version reaches the end of support, cluster control planes and nodes are automatically upgraded.
We recommend that if you want to prevent automatic upgrades for the entire Standard cluster or all of its node pools, use a maintenance exclusion with your cluster enrolled in a release channel. With maintenance exclusions, you can disable node auto-upgrades for all node pools, whereas you can disable node auto-upgrades at the node pool level if your cluster is not enrolled in a release channel.
Comparison between clusters enrolled and not enrolled in a release channel
Review the following table to understand the similarities and differences between enrolling and not enrolling your cluster in a release channel:
| Feature | Cluster enrolled in a release channel | Cluster not enrolled in a release channel | Shared upgrade behavior |
|
|---|---|---|
| Upgrade timing | Aligned with the respective release channel |
|
Control of node pool disruption |
|
|
| Maintenance windows | Available | Available |
| Maintenance exclusions |
Available maintenance exclusion scopes:
|
Restricted to "No upgrades" scope (30 days) |
| Rollout sequencing | Available with fleet-based and scope-based sequences | Not available |
| Long-term support | Available with the Extended release channel only | Not available |
| Autopilot | Available | Not available |
Differences between Rapid-channel clusters and alpha clusters
Clusters created using the Rapid release channel are not alpha clusters. Here are the differences:
- Clusters that use release channels can be upgraded, and auto-upgrade is enabled and cannot be disabled. Alpha clusters cannot be upgraded.
- Clusters that use release channels don't expire. Alpha clusters expire after 30 days.
- Alpha Kubernetes APIs are not enabled on clusters that use release channels.
What's next
- Use release channels
- Learn more about upgrading clusters.
- Get visibility into cluster upgrades (Preview)
- Learn how to Receive cluster upgrade notifications.
- Learn about managing automatic cluster upgrades across environments with rollout sequencing.