I am using Deployment kind to define a pod. Pod definition contains two containers and 1 sidecar.
In kubernetes events, Execution of container start is in sequential way. How we can start both containers parallely.
Container 1 (6s) + Container (5s) + NEG ready(2s) = 13s
If we start container in parallel we can minimize the pod spinning total time.
In kubernetes events, Execution of container start is in sequential way. How we can start both containers parallely.
Although container is started in sequence, kubelet does not wait for the container to enter running state to start the next container. The time gap for kubelet to loop thru is only fraction of second. It is the startup and readiness time that your containers needed that define how fast it can start receiving requests.
Containers within a pod are started in parallel, see this article for reference.
Perhaps there is some implicit dependency between your containers that is preventing them from starting at the same time.
