📓 Summary

OTel logs fail to validate when users modify fields that have been marked with time_series_dimension: true in the built-in component templates. The shared OTel component template applies time series dimension settings to resource.attributes and scope.attributes for both logs and metrics, but this causes a "shadowing" error when fields are modified in logs indices.

When a dynamic field like resource.attributes.service.name gets concretized to a keyword field during schema processing, Elasticsearch rejects it because it conflicts with the parent's time_series_dimension: true marker.

✔️ Acceptance Criteria

• time_series_dimension settings are ignored or not applied to non-TSDB indices (logs, events, etc.)
• Users can modify OTel log fields without encountering mapper_parsing_exception for shadowing violations on time_series_dimension
• Metrics indices continue to correctly apply time_series_dimension to resource.attributes as routing dimensions

🤓 Implementation Details

The OTel templates need differentiation between logs and metrics:

• Current state: Both logs-otel@mappings and metrics-otel@mappings share the same otel@mappings component template which marks attributes with time_series_dimension: true
• Expected behavior: The dimension definition should only apply to time series data streams
• Solution approach: Ensure that Elasticsearch ignores time_series_dimension constraints for indices that are not in TSDB mode, or create separate component templates for logs vs metrics that respect their respective index modes