[Remote translog] Auto restore an index if it has been in red state for atleast x minutes/seconds #6736
Description
Is your feature request related to a problem? Please describe.
For remote-backed indexes (both translog and segments are stored remotely), request level durability is available. Any request that has been acknowledged can be restored even if there are node failures. For remote-backed indexes, we already have API capability to restore an index manually upto the last acknowledged write.
Describe the solution you'd like
Changes need to be done on Cluster Manager side to close a red index and trigger auto restore.
In future, we need to be able to only restore the shards that are red instead of restoring the entire index.