[Ingest Manager] Fleet: Gracefully handle overloaded Elasticsearch

[roncohen]

In fleet routes, we've seen some 503s and 409s happening when Elasticsearch is overloaded. It would be good confirm that this is the cause of the 503s and 409s. If that turns out to be right, we need to work towards replying to clients with 429 in those cases to indicate that fleet overloaded at the moment to allow client to exponentially back off.

There has also been new developments on the Elasticsearch side that we should be tracking to make sure we handle gracefully.

ping @ph @nchaulet

[elasticmachine]

Pinging @elastic/ingest-management (Team:Ingest Management)

[elasticmachine]

Pinging @elastic/ingest-management (Feature:Fleet)

[ph]

Have we done any progress on this story in the scaling work?

[ph]

going to close this as we move to the go server implementation.