Feature request: wait for resources to exist. #8
Comments
|
Could you provide the specific script that would reproduce this issue? I think it makes sense, but not sure I exactly understand the mechanism that causes kube-health to not wait properly. |
I'm expecting deployment/foo to be re-created by an operator and I want to wait till it exists and is healthy. However kube-health will exit immediately if the deployment does not exist. So I want to add "wait-for-exists" before the health check. Perhaps a better API:
That would give the extra feature of being able to wait just for existence by using -E on its own. |
|
The less flags the better, having there some grace period, especially when waiting for the state, is a good idea IMO |
Use case: I delete a resource to force it to be re-created by an operator.
I want to use kube-health to wait until the resource is re-created and progresses to a healthy state.
Current behavior:
kube-health -Owith a non-existent resource name returns immediately.When called immediately after
kubectl deleteit usually does not wait because the resource does not yet exist, as there is an unpredicatble delay until the operator re-creates it.Desired behavior: with waiting flags (-W or -O) kube-health first waits for the resource to exist, then waits for it to progress.
Backwards compatibility: add
--ignore-not-foundflag to return immediately with no error if resource not found (current behavior).This is consistent with the meaning of --ignore-not-found as used by kubectl.
The text was updated successfully, but these errors were encountered: