Fix Get trained models statistics API types

[svalbuena]

Fixes #2762, I've tested these changes by downloading the .net client project, doing the equivalent changes done in this spec PR, and using it in the project I'm working on to invoke the GET http://localhost:9200/_ml/trained_models/intfloat__e5-small-v2/_stats endpoint, I now get a successful response.

Sample response of GET http://localhost:9200/_ml/trained_models/intfloat__e5-small-v2/_stats

{
    "count": 1,
    "trained_model_stats": [
        {
            "model_id": "intfloat__e5-small-v2",
            "model_size_stats": {
                "model_size_bytes": 132921240,
                "required_native_memory_bytes": 734682164
            },
            "pipeline_count": 1,
            "ingest": {
                "total": {
                    "count": 0,
                    "time_in_millis": 0,
                    "current": 0,
                    "failed": 0
                },
                "pipelines": {
                    "custom_embedding": {
                        "count": 0,
                        "time_in_millis": 0,
                        "current": 0,
                        "failed": 0,
                        "processors": [
                            {
                                "script": {
                                    "type": "script",
                                    "stats": {
                                        "count": 0,
                                        "time_in_millis": 0,
                                        "current": 0,
                                        "failed": 0
                                    }
                                }
                            },
                            {
                                "foreach": {
                                    "type": "foreach",
                                    "stats": {
                                        "count": 0,
                                        "time_in_millis": 0,
                                        "current": 0,
                                        "failed": 0
                                    }
                                }
                            }
                        ]
                    }
                }
            },
            "inference_stats": {
                "failure_count": 0,
                "inference_count": 13,
                "cache_miss_count": 0,
                "missing_all_fields_count": 0,
                "timestamp": 1722591230386
            },
            "deployment_stats": {
                "deployment_id": "intfloat__e5-small-v2",
                "model_id": "intfloat__e5-small-v2",
                "threads_per_allocation": 1,
                "number_of_allocations": 1,
                "queue_capacity": 1024,
                "state": "started",
                "allocation_status": {
                    "allocation_count": 1,
                    "target_allocation_count": 1,
                    "state": "fully_allocated"
                },
                "cache_size": "126.7mb",
                "priority": "normal",
                "start_time": 1722254553506,
                "inference_count": 13,
                "peak_throughput_per_minute": 5,
                "nodes": [
                    {
                        "node": {
                            "u8L8FqPtSb-eBsbz3yroew": {
                                "name": "d2acd1f5a58c",
                                "ephemeral_id": "q6hDvPcpQJaab7lSOE4sZw",
                                "transport_address": "172.18.0.2:9300",
                                "external_id": "d2acd1f5a58c",
                                "attributes": {
                                    "ml.max_jvm_size": "1073741824",
                                    "ml.allocated_processors_double": "16.0",
                                    "ml.allocated_processors": "16",
                                    "ml.machine_memory": "56860405760",
                                    "transform.config_version": "10.0.0",
                                    "xpack.installed": "true",
                                    "ml.config_version": "12.0.0"
                                },
                                "roles": [
                                    "data",
                                    "data_cold",
                                    "data_content",
                                    "data_frozen",
                                    "data_hot",
                                    "data_warm",
                                    "ingest",
                                    "master",
                                    "ml",
                                    "remote_cluster_client",
                                    "transform"
                                ],
                                "version": "8.12.0",
                                "min_index_version": 7000099,
                                "max_index_version": 8500008
                            }
                        },
                        "routing_state": {
                            "routing_state": "started"
                        },
                        "inference_count": 13,
                        "average_inference_time_ms": 14.615384615384615,
                        "average_inference_time_ms_excluding_cache_hits": 95,
                        "inference_cache_hit_count": 11,
                        "last_access": 1722591142606,
                        "number_of_pending_requests": 0,
                        "start_time": 1722588669854,
                        "threads_per_allocation": 1,
                        "number_of_allocations": 1,
                        "peak_throughput_per_minute": 5,
                        "throughput_last_minute": 2,
                        "average_inference_time_ms_last_minute": 0,
                        "inference_cache_hit_count_last_minute": 2
                    }
                ]
            }
        }
    ]
}

Closes #2770

[flobernd]

Hi @svalbuena, thanks again for your contribution 🙂 I left some comments.

[flobernd]

I think DateTime might have been correct here, as it's defined as:

/**
 * A date and time, either as a string whose format can depend on the context (defaulting to ISO 8601), or a
 * number of milliseconds since the Epoch. Elasticsearch accepts both as input, but will generally output a string
 * representation.
 */
export type DateTime = string | EpochTime<UnitMillis>

We have to check the server-code in order to verify if it should be DateTime or just EpochTime<UnitMillis>.

In any way:

The .NET client correctly uses DateTimeOffset as the CLR type, but fails to deserialize it from the epoch-time representation. This is something that must be fixed on the client itself and not in the spec.

[svalbuena]

interesting, in this case it seems EpochTime<UnitMillis> is the right type. I've checked other classes that use the same kind of value like start_time_in_millis of AsyncSearchResponseBase and it uses EpochTime for the one that is long. (I don't find any example of the start_timefield in the tests I've done nor in https://www.elastic.co/guide/en/elasticsearch/reference/current/async-search.html).

I'll set the field to be EpochTime for now until you find the right answer or fix for the .net client.

[flobernd]

Yes, this seems right. I think this should be the corresponding server code:
https://github.com/elastic/elasticsearch/blob/02c494963a59610a0be07c7d54337017a1e5beaf/x-pack/plugin/ml/src/main/java/org/elasticsearch/xpack/ml/inference/TrainedModelStatsService.java#L282C58-L282C106

[svalbuena]

so EpochTime<UnitMillis> looks good? anything else to address in this PR?

[flobernd]

LGTM from my side, thank you!

Let's wait for a second pair of eyes to double-check it.

[l-trotta]

LGTM as well!

[svalbuena]

How can we merge this? @l-trotta @flobernd

[pquentin]

@svalbuena Since validation only runs from branches, I opened #2770 with your changes, which allowed validation to pass. Your pull request did not affect the reported failures, since apparently tests did not surface this specific issue. (ML request issues are tracked in #2621.)

Thanks, merging!