Skip to content

Commit 9596d90

Browse files
Merge pull request #222562 from allyford/cluster-cleanup-policy
Update stopped cluster policy
2 parents 16000aa + 1afcade commit 9596d90

File tree

1 file changed

+5
-5
lines changed

1 file changed

+5
-5
lines changed

articles/aks/support-policies.md

Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -127,17 +127,17 @@ Although you can sign in to and change agent nodes, doing this operation is disc
127127

128128
You may only customize the NSGs on custom subnets. You may not customize NSGs on managed subnets or at the NIC level of the agent nodes. AKS has egress requirements to specific endpoints, to control egress and ensure the necessary connectivity, see [limit egress traffic](limit-egress-traffic.md). For ingress, the requirements are based on the applications you have deployed to cluster.
129129

130-
## Stopped or de-allocated clusters
130+
## Stopped, de-allocated, and "Not Ready" nodes
131131

132-
As stated earlier, manually de-allocating all cluster nodes via the IaaS APIs/CLI/portal renders the cluster out of support. The only supported way to stop/de-allocate all nodes is to [stop the AKS cluster](start-stop-cluster.md#stop-an-aks-cluster), which preserves the cluster state for up to 12 months.
132+
If you do not need your AKS workloads to run continuously, you can [stop the AKS cluster](start-stop-cluster.md#stop-an-aks-cluster) which stops all nodepools and the control plane, and start it again when needed. When you stop a cluster using the `az aks stop` command, the cluster state will be preserved for up to 12 months. After 12 months the cluster state and all of its resources will be deleted.
133133

134-
Clusters that are stopped for more than 12 months will no longer preserve state.
134+
Manually de-allocating all cluster nodes via the IaaS APIs/CLI/portal is not a supported way to stop an AKS cluster or nodepool. The cluster will be considered out of support and will be stopped by AKS after 30 days. The clusters will then be subject to the same 12 month preservation policy as a correctly stopped cluster.
135135

136-
Clusters that are de-allocated outside of the AKS APIs have no state preservation guarantees. The control planes for clusters in this state will be archived after 30 days, and deleted after 12 months.
136+
Clusters with 0 "Ready" nodes (or all "Not Ready") and 0 Running VMs will be stopped after 30 days.
137137

138138
AKS reserves the right to archive control planes that have been configured out of support guidelines for extended periods equal to and beyond 30 days. AKS maintains backups of cluster etcd metadata and can readily reallocate the cluster. This reallocation can be initiated by any PUT operation bringing the cluster back into support, such as an upgrade or scale to active agent nodes.
139139

140-
If your subscription is suspended or deleted, your cluster's control plane and state will be deleted after 90 days.
140+
All clusters in a suspended or deleted subscription will be stopped immediately and deleted after 30 days
141141

142142
## Unsupported alpha and beta Kubernetes features
143143

0 commit comments

Comments
 (0)