Support CCS minimize round trips in async search

[quux00]

This commit makes the smallest set of changes to allow async-search based cross-cluster search
to work with the CCS minimize_round_trips feature without changing the internals/architecture of
the search action.

When ccsMinimizeRoundtrips is set to true on SubmitAsyncSearchRequest, the AsyncSearchTask on the
primary CCS coordinator sends a synchronous SearchRequest to all to clusters for a remote coordinator
to orchestrate and return the entire result set to the CCS coordinator as a single response.

This is the same functionality provided by synchronous CCS search using minimize_round_trips.
However, since this is an async search, it means that the async search coordinator has no visibility
into search progress on the remote clusters while they are running the search, thus losing one of
the key features of async search. However, this is a good first approach for improving overall search
latency for cross cluster searches that query a large number of shards on remote clusters, since
Kibana does not currently expose incremental progress of an async search to users.

[elasticsearchmachine]

Pinging @elastic/es-search (Team:Search)

[javanna]

looks good to me so far, I am curious on why it's marked WIP ;) maybe we should add some indication in the get async search API response that some remote clusters are minimizing roundtrips?

[quux00]

With these changes, I have removed the notifyMinimizeRoundtrips/onMinimizeRoundtrips call to the SearchProgressListener in favor of modifying the SearchResponse.Cluster object to hold the information needed.

The _clusters object in the REST response for async_search with minimizeRoundTrips starts out at "successful: 0" and can move to "successful: 1" if there is a local cluster search that finishes while one or more of the remote clusters are still doing the search.

Here's an example of the progression in the _clusters output for a CCS search using minimize-roundtrips against three 3-node clusters:

POST blogs,remote1:*,remote2:*/_async_search?ccs_minimize_roundtrips=true { // ... search body for a slow search not shown } // after _async_seach submit { "id": "FkF3RE1KZnFSU3M2VUdsWEZjVnlmZFEaQVRnZDdaNjBUZ1N0RU1sUjl0QmphUTo2OTI=", "is_partial": true, "is_running": true, "start_time_in_millis": 1684516160099, "expiration_time_in_millis": 1684948160099, "response": { "took": 806, "timed_out": false, "terminated_early": false, "num_reduce_phases": 0, "_shards": { "total": 3, // only shows local shards until remote searches finish (current behavior) "successful": 0, // no shards completed "skipped": 0, "failed": 0 }, "_clusters": { "total": 3, "successful": 0, // no clusters completed (new behavior) "skipped": 0 }, "hits": { "total": { "value": 0, "relation": "gte" }, "max_score": null, "hits": [] } } } // while running (local cluster has finished; two remote cluster searches still running GET /_async_search/FkF3RE1KZnFSU3M2VUdsWEZjVnlmZFEaQVRnZDdaNjBUZ1N0RU1sUjl0QmphUTo2OTI { "id": "FkF3RE1KZnFSU3M2VUdsWEZjVnlmZFEaQVRnZDdaNjBUZ1N0RU1sUjl0QmphUTo2OTI=", "is_partial": true, "is_running": true, "start_time_in_millis": 1684516160099, "expiration_time_in_millis": 1684948160099, "response": { "took": 5973, "timed_out": false, "terminated_early": false, "_shards": { "total": 3, "successful": 3, // local shards completed "skipped": 0, "failed": 0 }, "_clusters": { "total": 3, "successful": 1, // local cluster has completed (new behavior) "skipped": 0 }, "hits": { "total": { "value": 142, "relation": "eq" }, "max_score": null, "hits": [] } } } // when all clusters completed GET /_async_search/FkF3RE1KZnFSU3M2VUdsWEZjVnlmZFEaQVRnZDdaNjBUZ1N0RU1sUjl0QmphUTo2OTI { "id": "FkF3RE1KZnFSU3M2VUdsWEZjVnlmZFEaQVRnZDdaNjBUZ1N0RU1sUjl0QmphUTo2OTI=", "is_partial": false, "is_running": false, "start_time_in_millis": 1684516160099, "expiration_time_in_millis": 1684948160099, "response": { "took": 10350, "timed_out": false, "num_reduce_phases": 4, "_shards": { "total": 9, "successful": 9, "skipped": 0, "failed": 0 }, "_clusters": { "total": 3, "successful": 3, "skipped": 0 }, "hits": { "total": { "value": 432, "relation": "eq" }, "max_score": 17.110811, "hits": [ ...

Note that this is NOT the behavior with ccs_minimize_roundtrips=false. There there _clusters output is always the final output where total == successful + skipped:

POST blogs,remote1:*,remote2:*/_async_search?ccs_minimize_roundtrips=false { // ... search body for a slow search not shown } ... "id": "FkF3RE1KZnFSU3M2VUdsWEZjVnlmZFEaQVRnZDdaNjBUZ1N0RU1sUjl0QmphUTo2OTI=", "is_partial": true, "is_running": true, "start_time_in_millis": 1684516160099, "expiration_time_in_millis": 1684948160099, "response": { ... "_clusters": { "total": 3, "successful": 3, // shows final state even if no shards/clusters have completed (current behavior) "skipped": 0 },

[quux00]

I tested a case where the remote clusters are down:

// case where skip_unavailable: true is NOT set for a remote cluster and it is down: POST blogs,remote1:*,remote2:*/_async_search?ccs_minimize_roundtrips=true ... // response: { "id": "FjJJZG9yWUhOU3VLaTJzNXZMalZaTkEaaVhraVNDT2VTenFoUDlod21hYmlKZzo5MTE=", "is_partial": true, "is_running": false, "start_time_in_millis": 1684767109136, "expiration_time_in_millis": 1685199109136, "response": { "took": 10927, "timed_out": false, "terminated_early": false, "_shards": { "total": 3, "successful": 3, "skipped": 0, "failed": 0 }, "_clusters": { "total": 3, "successful": 1, "skipped": 0 }, "hits": { "total": { "value": 167, "relation": "eq" }, "max_score": null, "hits": [] } }, "error": { "type": "status_exception", "reason": "error while executing search", "caused_by": { "type": "connect_transport_exception", "reason": "[][127.0.0.1:9301] connect_exception", "caused_by": { "type": "uncategorized_execution_exception", "reason": "Failed execution", "caused_by": { "type": "execution_exception", "reason": "execution_exception: io.netty.channel.AbstractChannel$AnnotatedConnectException: Connection refused: /127.0.0.1:9301", "caused_by": { "type": "i_o_exception", "reason": "Connection refused: /127.0.0.1:9301", "caused_by": { "type": "i_o_exception", "reason": "Connection refused" } } } } } } } ---- // where skip_unavailable:true POST blogs,remote1:*,remote2:*/_async_search?ccs_minimize_roundtrips=true ... // response: { "id": "FnE0SjZrb2lOVFR5ejFCN3dtX0NHbVEbaVhraVNDT2VTenFoUDlod21hYmlKZzoxMDIz", "is_partial": false, "is_running": false, "start_time_in_millis": 1684767263845, "expiration_time_in_millis": 1685199263845, "response": { "took": 11943, "timed_out": false, "num_reduce_phases": 3, "_shards": { "total": 6, "successful": 6, "skipped": 0, "failed": 0 }, "_clusters": { "total": 3, "successful": 2, "skipped": 1 // was properly updated at the end (started off as 0) }, "hits": { "total": { "value": 370, "relation": "eq" }, "max_score": 16.595804, "hits": [ { ...

[javanna]

left a couple of comments, great progress!!

[javanna]

SHIP IT

[quux00]

@abdonpijpelink - Would you have time to review the end user docs changes here?

You can look at the last commit to see the new changes: a7b7237

[quux00]

@abdonpijpelink

I just talked with Luca and we are going to change the default setting in a follow-on PR, so the end user docs changes will go into that new PR so you can wait to review it then. I'll be deleting those commits from this PR in a moment. Sorry for the change.

[abdonpijpelink]

Sure thing! For your follow-up PR: note that the <1> numbers ("callouts") have to restart from 1 for each code snippet, otherwise the docs won't build.

[abdonpijpelink]

Suggested change

	"successful": 3, <4>
	"skipped": 0,
	"failed": 0
	},
	"_clusters": {
	"total": 3,
	"successful": 1, <5>
	"skipped": 0
	},
	"hits": {
	"total": {
	"value": 167, <6>
	"relation": "eq"
	},
	"max_score": null,
	"hits": []
	}
	}
	}
	--------------------------------------------------
	<4> All the local cluster shards have completed.
	<5> The local cluster search has completed, so the "successful" clusters entry
	is set to 1. The `_clusters` response section will not be updated for the remote clusters
	until all remote searches have finished (either successfully or been skipped).
	<6> Number of hits from the local cluster search. Final hits are not
	"successful": 3, <1>
	"skipped": 0,
	"failed": 0
	},
	"_clusters": {
	"total": 3,
	"successful": 1, <2>
	"skipped": 0
	},
	"hits": {
	"total": {
	"value": 167, <3>
	"relation": "eq"
	},
	"max_score": null,
	"hits": []
	}
	}
	}
	--------------------------------------------------
	<1> All the local cluster shards have completed.
	<2> The local cluster search has completed, so the "successful" clusters entry
	is set to 1. The `_clusters` response section will not be updated for the remote clusters
	until all remote searches have finished (either successfully or been skipped).
	<3> Number of hits from the local cluster search. Final hits are not