[BUG] type param not working as expected for the Top N queries API

[deshsidd]

What is the bug?

Need to filter out memory and cpu when we query top queries using param :
GET /_insights/top_queries?type=latency
However, looks like this is not getting filtered out and might be a regressions.

How can one reproduce the bug?

Enabled topN:

{
  "persistent" : {
    "search.insights.top_queries.latency.window_size" : "1m",
    "search.insights.top_queries.latency.top_n_size" : 3,
    "search.insights.top_queries.latency.enabled" : true,
    "search.insights.top_queries.cpu.enabled" : true,
    "search.insights.top_queries.memory.enabled" : true
  }
}

Index docs and run a few search queries.

GET http://{{hostname}}/_insights/top_queries?type=latency

{
            "timestamp": 1724887236541,
            "search_type": "query_then_fetch",
            "indices": [
                "my_index"
            ],
            "task_resource_usages": [
                {
                    "action": "indices:data/read/search[phase/query]",
                    "taskId": 51,
                    "parentTaskId": 50,
                    "nodeId": "CFpm76yZQ5iecw8ciQFV3g",
                    "taskResourceUsage": {
                        "cpu_time_in_nanos": 4063000,
                        "memory_in_bytes": 33656
                    }
                },
                {
                    "action": "indices:data/read/search",
                    "taskId": 50,
                    "parentTaskId": -1,
                    "nodeId": "CFpm76yZQ5iecw8ciQFV3g",
                    "taskResourceUsage": {
                        "cpu_time_in_nanos": 366000,
                        "memory_in_bytes": 3224
                    }
                }
            ],
            "phase_latency_map": {
                "expand": 0,
                "query": 14,
                "fetch": 1
            },
            "node_id": "CFpm76yZQ5iecw8ciQFV3g",
            "total_shards": 1,
            "source": {
                "query": {
                    "match_all": {
                        "boost": 1.0
                    }
                }
            },
            "labels": {},
            "latency": 20,
            "cpu": 4429000,
            "memory": 36880
        }

What is the expected behavior?

returns latency, cpu and memory but should only return latency. Same for all the other combinations.

What is your host/environment?

mac localhost

Do you have any screenshots?

If applicable, add screenshots to help explain your problem.

Do you have any additional context?

Add any other context about the problem.

[dzane17]

This behavior is only when search.query.metrics.enabled is true.

We append latency, cpu, memory data here based on shouldCollect() which checks the individual metric flags OR search.query.metrics.enabled:

    private boolean shouldCollect(MetricType metricType) {
        return queryInsightsService.isSearchQueryMetricsFeatureEnabled() || queryInsightsService.isCollectionEnabled(metricType);
    }

https://github.com/opensearch-project/query-insights/blob/main/src/main/java/org/opensearch/plugin/insights/core/listener/QueryInsightsListener.java#L175-L177

[dzane17]

When search.query.metrics.enabled=false, top N only returns metrics which are enabled.

Example: 1st query has only latency, 2nd query has latency & cpu

curl -X GET "localhost:9200/_insights/top_queries?pretty"                                             
{
  "top_queries" : [
    {
      "timestamp" : 1724957602437,
      "phase_latency_map" : {
        "expand" : 0,
        "query" : 1,
        "fetch" : 0
      },
      "labels" : { },
      "source" : {
        "query" : {
          "match_all" : {
            "boost" : 1.0
          }
        }
      },
      "indices" : [ ],
      "search_type" : "query_then_fetch",
      "node_id" : "DCHfS-bKQHyLsBcqH5j6MQ",
      "task_resource_usages" : [
        {
          "action" : "indices:data/read/search[phase/query]",
          "taskId" : 265,
          "parentTaskId" : 264,
          "nodeId" : "DCHfS-bKQHyLsBcqH5j6MQ",
          "taskResourceUsage" : {
            "cpu_time_in_nanos" : 811000,
            "memory_in_bytes" : 18248
          }
        },
        {
          "action" : "indices:data/read/search",
          "taskId" : 264,
          "parentTaskId" : -1,
          "nodeId" : "DCHfS-bKQHyLsBcqH5j6MQ",
          "taskResourceUsage" : {
            "cpu_time_in_nanos" : 280000,
            "memory_in_bytes" : 3272
          }
        }
      ],
      "total_shards" : 1,
      "latency" : 3
    },
    {
      "timestamp" : 1724957554669,
      "phase_latency_map" : {
        "expand" : 0,
        "query" : 2,
        "fetch" : 0
      },
      "labels" : { },
      "source" : {
        "query" : {
          "match_all" : {
            "boost" : 1.0
          }
        }
      },
      "indices" : [ ],
      "search_type" : "query_then_fetch",
      "node_id" : "DCHfS-bKQHyLsBcqH5j6MQ",
      "task_resource_usages" : [
        {
          "action" : "indices:data/read/search[phase/query]",
          "taskId" : 243,
          "parentTaskId" : 242,
          "nodeId" : "DCHfS-bKQHyLsBcqH5j6MQ",
          "taskResourceUsage" : {
            "cpu_time_in_nanos" : 793000,
            "memory_in_bytes" : 18248
          }
        },
        {
          "action" : "indices:data/read/search",
          "taskId" : 242,
          "parentTaskId" : -1,
          "nodeId" : "DCHfS-bKQHyLsBcqH5j6MQ",
          "taskResourceUsage" : {
            "cpu_time_in_nanos" : 182000,
            "memory_in_bytes" : 3272
          }
        }
      ],
      "total_shards" : 1,
      "cpu" : 975000,
      "latency" : 3
    }
  ]
}

[ansjcy]

This is an expected behavior, we always measure latency, cpu and memory usage unless you turn off search.query.metrics.enabled as @dzane17 memtioned.
when you do GET /_insights/top_queries?type=latency, it means "get the top queries sorted by latency", but all other metrics will be included, it's just we are not sorting based on those keys.