[elasticsearch] Add a way for the readiness check to accept yellow state temporarily

Describe the feature: Enable manually allowing the readiness check to accept yellow state temporarily. E.g. by creating a file inside the pod, which would be deleted automatically once it reaches the green state.

Describe a specific use case for the feature: When all the nodes crash and then are starting again, e.g. after depleting memory/heap on all nodes, all the nodes wait for the green state, even though the cluster is operational once it reaches the yellow state.

Updating the statefulset with clusterHealthCheckParams set without wait_for_green isn’t the best solution as it entails another restart of the whole cluster.

[jmlrt]

Hi @mtruneck, Thanks for submitting this request, unfortunately we won’t have time to think about it soon so please be patient.

[travnewmatic]

i want this too

i’m running a single node intentionally

yellow is acceptable for me.