Elasticsearch: Make Elasticsearch job to ignore service(s) spans (#92)

[mohammedalics]

Signed-off-by: Mohammed Ammer mohammed.ali.cs@gmail.com

Resolves #92

Which problem is this PR solving?

Make Elasticsearch job to ignore service(s) spans

Short description of the changes

• Introduce a new environment variable ES_SERVICES_IGNORE
• Update the Elasticsearch job query to consider the names of the services in ES_SERVICES_IGNORE.

The current implemented query is like (considering 2h spans from now)

{
  "query": {
    "range": {
      "startTimeMillis": {
        "gte": "now-2h"
      }
    }
  }
}

By considering ignoring services (considering 2h spans from now and ignoring serviceX and serviceY), the query will be like

{
  "query": {
    "bool": {
      "must_not": [
        {
          "match_phrase": {
            "process.serviceName": {
              "query": "serviceX"
            }
          }
        },
        {
          "match_phrase": {
            "process.serviceName": {
              "query": "serviceY"
            }
          }
        }
      ],
      "must": [
        {
          "range": {
            "startTimeMillis": {
              "gte": "now-2h"
            }
          }
        }
      ]
    }
  }
}

[pavolloffay]

Excluding the services in the query has certainly some advantages.

I am wondering whether it would be better to filter the spans in the generic job so it works for all storages.

[mohammedalics]

Good idea.

Would you prefer to consider it part of this PR or separate?

[pavolloffay]

It can be done in this PR as it is implementing it.