Indexing Threads chart in Node > Advanced tab is empty

[ycombinator]

Kibana version: master

Elasticsearch version: master

Server OS version: Mac OS X

Browser version: Chrome latest

Browser OS version: Mac OS X

Original install method (e.g. download page, yum, from source, etc.): From source

Describe the bug:

See Indexing Threads chart and corresponding node_index_threads field in response JSON in screenshot below:

Steps to reproduce:

1. Turn on monitoring
2. Visit the Elasticsearch Node Listing page
3. Click on a node
4. Visit the Advanced tab for the node

Expected behavior:

I would expect the chart to show some values, even if they are 0s. Similarly the chart's legend should some values other than N/A.

[ycombinator]

@chrisronline FYI, this is happening with internal ES collection as well (that is, no metricbeat collection in the picture).

[chrisronline]

The issue seems to be a lack of mappings for these fields. :(

[chrisronline]

We should probably do an inventory of fields used in the monitoring UI and compare them to what fields exist in the mappings and report on the missing ones

[chrisronline]

To clarify, this or implies that if any of these fields are present, it will work properly. I'm not sure why none are there though but we still need to fix the mappings.

@pickypg any thoughts here?

[pickypg]

I don't know what happened to the write threadpool mapping, but that's what is missing here since I assume you are testing against master.

If you add

PUT /.monitoring-es-*/_mapping/doc
{
  "doc": {
    "properties": {
      "node_stats": {
        "properties": {
          "thread_pool": {
            "properties": {
              "write": {
                "properties": {
                  "threads": {
                    "type": "integer"
                  },
                  "queue": {
                    "type": "integer"
                  },
                  "rejected": {
                    "type": "long"
                  }
                }
              }
            }
          }
        }
      }
    }
  }
}

then run

POST /.monitoring-es-*/_update_by_query
{
  "query": {
    "match_all": { }
  }
}

then I expect it will start working (and show earlier data). Naturally this won't fix tomorrow's index, but we can then push that fix into the spot @chrisronline pointed out.

[ycombinator]

@chrisronline This issue can be closed now because it was resolved in elastic/elasticsearch#34392, correct?

[chrisronline]

@ycombinator Yup, good call!