Add OpenSearch Sink option to include Event tags in the document

[kkondaka]

Description

Add OpenSearch Sink option to include Event tags in the document.
This change allows writing events tags to the event document in the opensearch. A new opensearch config option tags_key_name is added. When this option is present, the tags_key_name value is used as name of the key to store the tags in the document.

Example config

- opensearch:                                                                                                                                          
       hosts: ["https://opensearch-node1:9200"]                                                                                                            
       tags_key_name: "tags"                                                                                                                                                                                                                                                                
       index: test_index

Example output

     {
        "_index": "new_http_logs_may_31",
        "_id": "nugsdYgBRCBJrQoAlWPJ",
        "_score": 1,
        "_source": {
          "message": "wordone wordtwo wordthree",
          "tags": [
            "tag1",
            "tag2"
          ]
        }
      }

Resolves #629

Issues Resolved

Resolves #629

Check List

• [X ] New functionality includes testing.
• [X ] New functionality has been documented.
  • New functionality has javadoc added
• [X ] Commits are signed with a real name per the DCO

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[graytaylor0]

I think we should stick with my proposal here (#629 (comment)). All of the sinks need tag support, and we can provide a common config that all sinks can add.

Adding just the tags_key_name here is not ideal for 2 reasons.

1. It doesn't consolidate the tagging parameter names (we will go back to adding this key name manually to all the sink plugins, and they could stop being consistent).
2. It doesn't scale very well if we want to add features/options for how the sink handles tags. I think the tagging config in the link above is very direct and easy to understand.

I think we at least need to consolidate whatever configuration / parameter we choose so that if it changes, we don't have to change code in every sink plugin. We could do this a number of ways I think. One option is to create an annotation in data-prepper-api (something like @DataPrepperSinkTagsKey that will set the @JsonProperty("tags_key_name")

[dlvenable]

We did something similar with routes which made the routes available as a global property.

Here is some of the work for it.

data-prepper/data-prepper-api/src/main/java/org/opensearch/dataprepper/model/configuration/SinkModel.java

Lines 69 to 71 in 59c48e2

	@JsonInclude(JsonInclude.Include.NON_EMPTY)
	@JsonProperty("routes")
	private final List<String> routes;

I think that we could have Data Prepper core read this flag and then update the underlying Event for the sink. Then sinks don't have to have any tagging-specific code. Thoughts?

[kkondaka]

Thanks @dlvenable and @graytaylor0. It makes sense to make this a sink level JsonProperty. I will make it similar to routes sink option.

[kkondaka]

@dlvenable reading the flag can be done in the Data Prepper core but the flag needs to be set in each Sink separately because not all sinks will have this option set. Secondly, we cannot update underlying event (even the current approach I have here is incorrect) because modifying an event would automatically modify the event FOR all sinks because we do not make a copy today when there are non-pipeline-connector sinks.

Only way to do it without copying is to do event.toJsonStringWithTags() instead of event.toJsonString() in DocumentBuilder.build in OpenSearchSink (and equivalent code in other sinks)