Can I use nested attributes on aggregation in a visualization?

Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):

v.2.19.1 opensearch and relevant dashboards on Kubernetes using the official images.

Describe the issue:

: I have 2 clusters v2.19.1, on both cases with an index with a ‘json’ nested dynamic attribute which on ingestion has some logic with if .message looks like json: .json = decode_json(.message) (oversimplification of it), in both cases those are datastreams.In both cases after significant ingestion both index patterns are updated and the relevant json fields are available in discover, so it looks like it’s working.The differences start when we reach visualizations: In one cluster it works and I can split series by a json attribute but not on the second cluster (it does not show the dynamic json attributes).

Configuration:

I don’t see any relevant configuration. In the case of those indexes, both are datastreams with a json nested and dynamic attributes. Not all documents have this particular attribute.

Relevant Logs or Screenshots: