AWS OpenSearch Serverless - Bulk API - "Document ID is not supported in create/index operation request"

There seems to be conflicting documentation relating to OpenSearch's bulk API, specifically related to "Create" and "Index" actions.

Error

If I run a simple bulk request with the following body:

{"index": {"_id": "63a3b4074244c5b760010f1f", "_index": "index-client"}}
{"message": "My log message"}

I get the following error in the response:

{
  'items': [
    {
      'index': {
        'status': 400,
        '_id': '63a3b4074244c5b760010f1f',
        'error': {
          'reason': 'Document ID is not supported in create/index operation request',
          'type': 'illegal_argument_exception'
        },
        '_index': 'index-client'
      }
    }
  ],
  'errors': True,
  'took': 0
}

Question

Everything I can find (see docs below for reference) suggests that the Bulk API for the supported 2.0.x version of OpenSearch looks like the following:

POST _bulk
{ "delete": { "_index": "movies", "_id": "tt2229499" } }
{ "index": { "_index": "movies", "_id": "tt1979320" } }
{ "title": "Rush", "year": 2013 }
{ "create": { "_index": "movies", "_id": "tt1392214" } }
{ "title": "Prisoners", "year": 2013 }
{ "update": { "_index": "movies", "_id": "tt0816711" } }
{ "doc" : { "title": "World War Z" } }

with the Newline Delimited JSON (see https://www.ndjson.org) format separating actions and documents. If we focus on the "Create" and "Index" actions, these allow you to specify an "_id" field in the action. For "Index" actions, this will: "create a document if it doesn’t yet exist and replace the document if it already exists."

So, how is the "Index" action supposed to work if the action can't take an "_id" to correlate it with an existing document?

You might think that maybe it pulls the "_id" field from the document itself instead of from the action, but this will lead to the following error:

Field [_id] is a metadata field and cannot be added inside a document. Use the index API request parameters.

Googling "Document ID is not supported in create/index operation request" reveals nothing. I'm at a loss for how to do Bulk Index operations for AWS OpenSearch Serverless.

Documentation

OpenSearch docs

(Version 2.0 because "Serverless collections currently run OpenSearch version 2.0.x.")

• Bulk API: https://opensearch.org/docs/2.0/api-reference/document-apis/bulk/
• Intro to indexing: https://opensearch.org/docs/2.0/opensearch/index-data/#introduction-to-indexing
• Python clients' API: https://opensearch-project.github.io/opensearch-py/api-ref/clients/opensearch_client.html#opensearchpy.OpenSearch.bulk

AWS docs

• Mention of bulk API: https://docs.aws.amazon.com/opensearch-service/latest/developerguide/gsgupload-data.html#gsgmultiple-document
• Bulk api quick start: https://docs.aws.amazon.com/opensearch-service/latest/developerguide/quick-start.html#quick-start-bulk

Elastic docs

• Bulk API: https://www.elastic.co/guide/en/elasticsearch/reference/current/docs-bulk.html#bulk-api-request-body