Hello,
We are using Dynatrace in Application-Only Monitoring mode to monitor a Dockerized application.
Our deployment follows the official Dynatrace documentation:
https://docs.dynatrace.com/docs/ingest-from/setup-on-container-platforms/docker/set-up-oneagent-on-c...
1. Current configuration
The application is deployed using Docker Compose, with the following environment variable:
environment:
- OTEL_RESOURCE_ATTRIBUTES=primary_tags.host=1,primary_tags.env=RECETTE
The application is instrumented using the Dynatrace Application-Only Monitoring approach.
With this configuration, the primary_tags.host=1 and primary_tags.env=RECETTE attributes are correctly applied to the corresponding service in Dynatrace.
However, these tags are not available on the associated Process Group and Host entities.
2. Expected behavior
We would like to propagate the same tag values consistently across the related Dynatrace entities:
- Service: primary_tags.env=RECETTE
- Process Group: primary_tags.env=RECETTE
- Host / container entity: primary_tags.env=RECETTE
Our objective is to maintain consistent tagging across the application and its associated entities, enabling consistent filtering, segmentation, and analysis throughout Dynatrace.
The same requirement applies to primary_tags.host=1.
3. Questions
Could you please advise on the recommended and officially supported approach for this use case?
In particular:
- Ingest Rule Enrichment / Grail enrichment: Can an ingest enrichment rule be configured to propagate or derive the primary_tags.env value from OpenTelemetry resource attributes to the associated Process Group and Host entities? If so, which data source, matching conditions, and target attributes should be used?
- Primary Grail Tags: Can Primary Grail Tags be used to achieve consistent tagging across services, Process Groups, and Hosts in this Application-Only Monitoring setup? If not, what are their limitations in this context?
- Application-Only Monitoring limitations: Does this deployment mode support tagging the Process Group and the container/Host entity using OpenTelemetry resource attributes, or are these attributes limited to application-level entities and telemetry?
- Alternative recommended solution: If ingest enrichment is not suitable, would you recommend another supported approach, such as process-level tags, container metadata, a dedicated enrichment rule, or a different configuration?
- Configuration example: Could you provide a concrete configuration example for our Docker Compose deployment, including any required prerequisites or version constraints?
We would prefer a centralized, maintainable solution that avoids having to define the same tag values independently at multiple levels, if this is supported.
Regards
