[ML] Switch ModelRegistry to return ResourceyAlreadyExistsException instead of ElasticsearchStatusException

[jonathan-buttner]

This PR is to address an issue where we're seeing an increased number of error logs when storing persisting default inference endpoints.

The issue is that the exception that was returned in this PR: https://github.com/elastic/elasticsearch/pull/136569/files#diff-65fa96c525e72184c64209a9b6fd0f8130a7faf883464cbb21824996837c424cR674

Was causing an error to be logged here: https://github.com/elastic/elasticsearch/blob/main/x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/registry/ModelRegistry.java#L452 because we were no longer returning a ResourceAlreadyExistsException.

Testing

I can reproduce this by making multiple calls to GET _inference/_all at the same time on a newly created cluster. There should be an error log with something like:

[2025-11-03T10:01:47,736][ERROR][o.e.x.i.r.ModelRegistry  ] [runTask-0] Failed to store default inference id [.multilingual-e5-small-elasticsearch] org.elasticsearch.ElasticsearchStatusException: Inference endpoint [.multilingual-e5-small-elasticsearch] already exists

After the fix, that shouldn't exist anymore (it'll be at a debug level).

You should still see warnings though (I verified that these exist prior to my changes) if you send a bunch of the GET requests at the same time when a cluster is first started.

[2025-11-03T10:04:18,890][WARN ][o.e.x.i.r.ModelRegistry  ] [runTask-0] Failed to store document id: [model_.rerank-v1-elasticsearch] inference id: [.rerank-v1-elasticsearch] index: [.inference] bulk failure message [[.inference/ghwJ_yhfQ5OQ43-wgBRcZg][[.inference][0]] org.elasticsearch.index.engine.VersionConflictEngineException: [model_.rerank-v1-elasticsearch]: version conflict, document already exists (current version [1])]
[2025-11-03T10:04:18,893][WARN ][o.e.x.i.r.ModelRegistry  ] [runTask-0] Failed to store document id: [model_.rerank-v1-elasticsearch] inference id: [.rerank-v1-elasticsearch] index: [.secrets-inference] bulk failure message [[.secrets-inference/2FvlGp-SSyqIXVdm4TjKhQ][[.secrets-inference][0]] org.elasticsearch.index.engine.VersionConflictEngineException: [model_.rerank-v1-elasticsearch]: version conflict, document already exists (current version [1])]

[elasticsearchmachine]

Pinging @elastic/ml-core (Team:ML)

[jan-elastic]

LGTM