ingest: date_index_name processor template resolution

[jakelandis]

This change adds support for template snippet (e.g. {{foo}}) resolution
in the date_index_name processor. The following configuration options
will now resolve a templated value if so configured:

• index_name_prefix (e.g "index_name_prefix": "myindex-{{foo}}-")
• date_rounding (e.g. "date_rounding" : "{{bar}}")
• index_name_format (e.g."index_name_format": "{{baz}}")

Example:

POST _ingest/pipeline/_simulate
{
  "pipeline" : {
    "processors" : [
      { "date_index_name": 
        { "field": "@timestamp", 
          "index_name_prefix": "myindex-{{event_type}}-", 
          "date_rounding" : "{{index-period}}",
          "index_name_format": "{{index_format}}"
        }
      }
    ]
  },
  "docs" : [
     {"_source": {
       "@timestamp" : "2018-07-04T12:00:00.000Z",
       "event_type" : "rare",
       "index-period" : "y",
       "index_format": "yyyy"
     }}
    ]
}

results in

{
  "docs": [
    {
      "doc": {
        "_index": "<myindex-rare-{2018||/y{yyyy|UTC}}>",
        "_type": "_type",
        "_id": "_id",
        "_source": {
          "@timestamp": "2018-07-04T12:00:00.000Z",
          "event_type": "rare",
          "index-period": "y",
          "index_format": "yyyy"
        },
        "_ingest": {
          "timestamp": "2018-07-10T18:01:22.750869Z"
        }
      }
    }
  ]
}

edit updated example
edit2 updated example
edit3 removed note that is no longer relevant

[elasticmachine]

Pinging @elastic/es-core-infra

[martijnvg]

I think it is better that the indexNamePrefix, dateRounding and indexNameFormat become TemplateScript.Factory typed fields, like in other processor implementations. ValueSource is more meant to template generic values and in this case we know the templated values are always strings. Also then it is not needed to pass down ScriptService into the processor implementation and then the change in IngestDocument is not needed too.

[jakelandis]

@martijnvg - thanks for the review ! the changes requested are at f9a6ec1. This PR should be ready for re-review.

[jakelandis]

Jenkins test this please

[jakelandis]

Jenkins test this please

edit Issue logged for unrelated failure: #31948

[martijnvg]

LGTM 👍

[jakelandis]

@martijnvg - thanks! merged and will backport to 6.x.

edit: 6.x backport here: f19d83f