I have two containers in my Pod (one main app container and another one is just a simple nginx reverse proxy)
containers: - name: my-app - name: my-nginx lifecycle: preStop: exec: command: ["/bin/sleep", "30"] terminationGracePeriodSeconds: 40
and preStop hook to make sure app still has some time to finish received requests while the pod is being terminated.
My question is: will Kubernetes terminate
my-app container first completely ignoring
preStop hook on the
my-nginx container? And if so what should I do to deal with this issue? Add same hook to both containers?
Is there any other way to deal with zero-downtime rolling deployments? (To make sure no request is lost during the redeploy)