Hi everyone
So I need to upgrade my aks cluster to a later version
My plan is to
My pods are controlled by replicaset, state statefulset and daemonset
Is this the right approach for a mission critical workload?
If so What are other best conventions to follow to ensure smooth transition?
Thanks
My advice - do not do an upgrade on a mission critical cluster without dry-running it in a test cluster first.
Deploy a copy of what you have into another cluster at the version you are on and with the same nodepool setup. Do the upgrade there and repeat until you have ironed out any issues.
If you have only standard workloads as above + configmaps, secrets, etc you should be fine. Be aware of the kubernetes changelogs in case there are any APi version changes as you will need to address these, both in workloads themselves, or workloads you may have that depend on API versions of other resources, e.g. cluster autoscaler etc - which themselves may require upgrading.
Don’t just jump into the upgrade. Careful planning is required first.