Remove _nodes field from under cluster_stats as it's not being used

[ycombinator]

What it says on the tin.

Testing this PR

1. Pull down this PR and build Metricbeat.

   cd metricbeat
   mage build
   

2. Enable the elasticsearch-xpack Metricbeat module.

   ./metricbeat modules enable elasticsearch-xpack
   

3. Start Metricbeat

   ./metricbeat -e
   

4. Delete older .monitoring-es-* documents to start with a clean slate (makes it easier for testing).

   DELETE .monitoring-es-*
   

5. Using Kibana Console check that the Metricbeat-indexed monitoring documents for Elasticsearch don't contain the cluster_stats._nodes field.

   GET .monitoring-es-7-mb-*/_search?q=type:cluster_stats
   

[elasticmachine]

Pinging @elastic/stack-monitoring

[ycombinator]

jenkins, test this

[cachedout]

In testing this, I still see a _nodes field in cluster_stats:

          "cluster_stats" : {
            "cluster_uuid" : "Fz1ioHoiQyic2ovQ5cpT2w",
            "timestamp" : 1.556530486628E12,
            "status" : "yellow",
            "indices" : {
              "fielddata" : {
                "memory_size_in_bytes" : 12576,
                "evictions" : 0
              },
              "query_cache" : {
                "evictions" : 2084,
                "memory_size_in_bytes" : 4085183.0,
                "total_count" : 182204,
                "hit_count" : 10379,
                "miss_count" : 171825,
                "cache_size" : 1188,
                "cache_count" : 3272
              },
              "completion" : {
                "size_in_bytes" : 0
              },
              "segments" : {
                "count" : 99,
                "stored_fields_memory_in_bytes" : 86544,
                "norms_memory_in_bytes" : 13824,
                "index_writer_memory_in_bytes" : 1.6319468E7,
                "version_map_memory_in_bytes" : 568244,
                "doc_values_memory_in_bytes" : 454180,
                "fixed_bit_set_memory_in_bytes" : 53424,
                "term_vectors_memory_in_bytes" : 0,
                "max_unsafe_auto_id_timestamp" : -1,
                "file_sizes" : { },
                "memory_in_bytes" : 1283486.0,
                "terms_memory_in_bytes" : 663453,
                "points_memory_in_bytes" : 65485
              },
              "count" : 18,
              "shards" : {
                "replication" : 0,
                "index" : {
                  "shards" : {
                    "max" : 1,
                    "avg" : 1,
                    "min" : 1
                  },
                  "primaries" : {
                    "max" : 1,
                    "avg" : 1,
                    "min" : 1
                  },
                  "replication" : {
                    "min" : 0,
                    "max" : 0,
                    "avg" : 0
                  }
                },
                "total" : 18,
                "primaries" : 18
              },
              "docs" : {
                "count" : 440682,
                "deleted" : 124816
              },
              "store" : {
                "size_in_bytes" : 1.40293974E8
              }
            },
            "nodes" : {
              "network_types" : {
                "transport_types" : {
                  "security4" : 1
                },
                "http_types" : {
                  "security4" : 1
                }
              },
              "count" : {
                "data" : 1,
                "coordinating_only" : 0,
                "master" : 1,
                "ingest" : 1,
                "total" : 1
              },
              "jvm" : {
                "mem" : {
                  "heap_used_in_bytes" : 4.69675904E8,
                  "heap_max_in_bytes" : 1.037959168E9
                },
                "threads" : 141,
                "max_uptime_in_millis" : 1.24097021E8,
                "versions" : [
                  {
                    "vm_name" : "OpenJDK 64-Bit Server VM",
                    "vm_version" : "12+33",
                    "vm_vendor" : "Oracle Corporation",
                    "bundled_jdk" : true,
                    "using_bundled_jdk" : true,
                    "count" : 1,
                    "version" : "12"
                  }
                ]
              },
              "fs" : {
                "total_in_bytes" : 4.99963174912E11,
                "free_in_bytes" : 1.9303784448E11,
                "available_in_bytes" : 1.88011114496E11
              },
              "plugins" : [ ],
              "discovery_types" : {
                "zen" : 1
              },
              "packaging_types" : [
                {
                  "type" : "tar",
                  "count" : 1,
                  "flavor" : "default"
                }
              ],
              "versions" : [
                "8.0.0"
              ],
              "process" : {
                "cpu" : {
                  "percent" : 0
                },
                "open_file_descriptors" : {
                  "min" : 479,
                  "max" : 479,
                  "avg" : 479
                }
              },
              "os" : {
                "names" : [
                  {
                    "name" : "Mac OS X",
                    "count" : 1
                  }
                ],
                "pretty_names" : [
                  {
                    "pretty_name" : "Mac OS X",
                    "count" : 1
                  }
                ],
                "mem" : {
                  "total_in_bytes" : 3.4359738368E10,
                  "free_in_bytes" : 1.94225152E10,
                  "used_in_bytes" : 1.4937223168E10,
                  "free_percent" : 57,
                  "used_percent" : 43
                },
                "available_processors" : 12,
                "allocated_processors" : 12
              }
            },
            "_nodes" : {
              "failed" : 0,
              "total" : 1,
              "successful" : 1
            }

[ycombinator]

@cachedout I just tested this PR locally and I'm not seeing thecluster_stats._nodes field. Is it possible you were looking at an older document (indexed before this PR's code was run)? I added a step in the testing instruction to delete old .monitoring-es-* documents to start with a clean slate. Mind re-testing this PR with that step?

[cachedout]

@ycombinator Interesting. In my testing I explicitly deleted the index each time but perhaps I made a mistake somewhere? I will re-test.

[cachedout]

My mistake! This works as described.