Operate
This section covers working with a cluster after it's already running. Deploy decides how the cluster is built, and Configure turns its capabilities on or tunes them in vcluster.yaml.
Operate work splits into tasks you repeat and changes you plan ahead.
Routine tasks​
These are the tasks most cluster users come back to.
- Connect. Access and expose gets you a kubeconfig through the vCluster CLI or a service account, and exposes the cluster outside the control plane cluster.
- Pause idle clusters. Sleep compares manual pause with the two kinds of auto sleep, so you can pick one before you set it up.
- Apply configuration changes. Deploy config changes rolls an updated
vcluster.yamlout to a running cluster. - Back up and restore. Create snapshots saves a cluster's state, and Restore snapshots brings it back.
- Read logs. Logging explains the kinds of logs vCluster produces and where to find each one.
When something goes wrong, see Troubleshoot.
Planned changes​
Some changes to a running cluster are done infrequently and planned ahead. They include upgrading vCluster, rotating the CA certificate, and moving from ingress-nginx to Gateway API. For the backing store, you can upgrade etcd to v3.6 or migrate from deployed to embedded etcd. Clusters still on K3s must move to Kubernetes before upgrading to vCluster v0.33 or later.
Each has its own prerequisites, so read the page through before you start. Back up the cluster first.
Covered in other sections​
Tasks that change how a cluster is built, rather than how you run it, live in Deploy and Configure.
- To harden a cluster, see Secure vCluster deployments.
- To run the control plane with high availability, deploy it in high availability. For vCluster Standalone, see Standalone high availability.
- To size a control plane for scale, see Control plane sizing and performance.
- To change which resources sync between the cluster and the control plane cluster, see Resource synchronization.
- To open a
kubectlshell from the platform UI, see Kubectl shell in the vCluster Platform docs. - To add links such as a pull request to a cluster, see Custom links in the vCluster Platform docs.
- To monitor many clusters together, see Fleet observability in the vCluster Platform docs.