IHAC would like to enable cross-cluster search, but the docs limit that to clusters created after June 2020. How to understand the limitation? Is that something we can work around? Creating a new cluster for their 100B documents is going to be quite painful.
What a given autotune actually does? Right now it’s just lines on a graph, and sometimes the graph gets worse afterwards.
What would it take to get shard allocation control? Not being able to control shard routing has been a real painpoint for them when allocation goes wrong.
Similar, what’s the recommended way to kick a bad node from the cluster?