spinnaker
Spinnaker pipeline stages to do rolling restart with validation
Issue Summary:
I want build a spinnaker pipeline to restarts pods in a deployment for multiple cells(ex 50+ deployment workloads). Once restarted we need to validate all the rolling restarts pods successfully running and deployment workload status is ready. if all rolling restarts are successful it should pass the build else it should terminate.
as of know we are doing kubectl patch to do rolling restart of all the pods in a deployment, no validations to tell the rolling restart is successful. Please suggest good pipeline in spinnaker where we can do rolling restart and validation after restarts. I'm not finding any documents or example pipelines to achieve above rolling restart of pods and validation of deployment workloads once rolling restart is completed.
I think a stage for a rollout restart could be a useful feature. That only works with deployments though.
If you use a Patch Manifest stage it should wait for all the replicas to become available - is that what you mean by validation?
@mattgogerly Thanks for the suggestions. What if want to bring down all pods at time and launch all. any stage we can use.
for example we use kubectl scale --replica=0 manually to scale down and scale up again to take down pods at a time. what we can do this to automate. using Spinnaker
You should be able to use the Patch Manifest stage on the deployment/replicaset to patch the number of replicas to 0, and then another stage to patch it back to what you want?
This issue hasn't been updated in 45 days, so we are tagging it as 'stale'. If you want to remove this label, comment:
@spinnakerbot remove-label stale
This issue is tagged as 'stale' and hasn't been updated in 45 days, so we are tagging it as 'to-be-closed'. It will be closed in 45 days unless updates are made. If you want to remove this label, comment:
@spinnakerbot remove-label to-be-closed
This issue is tagged as 'to-be-closed' and hasn't been updated in 45 days, so we are closing it. You can always reopen this issue if needed.
