Overview

Software Catalog automatically discovers entities from Datadog Application Performance Monitoring (APM), Universal Service Monitoring (USM), and Real User Monitoring (RUM). You can also configure it to import entities from infrastructure metrics and logs.

Automatic discovery with APM, USM, and RUM

Datadog Software Catalog is pre-populated with entries detected through APM, eBPF-based autodiscovery with Universal Service Monitoring, and RUM applications.

All automatically detected entities appear in the Component Selector within Software Catalog.

APM and USM automatically detect the following entities types: service,datastore, queue, external providers, inferred services, and endpoints. APM SDKs identify dependencies of instrumented services and classify them as databases, queues, or third-party APIs—even if those dependencies are not directly instrumented. Custom instrumentation may affect how entities are discovered and how the service tag is assigned. To learn more, see APM Inferred Services.

RUM is responsible for discovering frontend apps entities.

For information about discovering endpoints, see Discovering Endpoints from APM.

Note: The performance metrics for an entity are tied to its primary operations. If a service is detected only by USM, the performance view shows USM metrics; if a entity is detected only by APM, or by both APM and USM, the performance view shows APM trace metrics.

Managing automatically-named services

To adjust how inferred entities appear in Software Catalog, you can:

• Create a renaming rule to identify an entity by a custom name.
• Filter out entities by type (database, queue, third-party).
• Remove service overrides such as service:my-service-http-client from your catalog or map.

Correlation requirements for APM

To ensure proper auto-correlation between APM and Software Catalog, follow these best practices:

• Ensure your APM SDK and Datadog Agent versions support inferred entities.
• Avoid arbitrary service overrides and use inferred entities instead.
  • Arbitrary service overrides can make correlation unreliable and clutter your catalog. They also make Software Catalog hard to use for those unfamiliar with the override names.
• Rename inferred entities, like datastores and queues, to match your team’s naming conventions.
• Use Unified Service Tagging to ensure service names, environments, and versions match across APM, infrastructure, and catalog definitions.
  • In Datadog APM, a service is an identifier used to group traces and aggregate metrics, not necessarily a logical service. Define what constitutes a “service” for your organization (for example, a Kubernetes namespace, deployment, or individual pod) and apply this definition consistently across all telemetry sources.

Import entities from Infrastructure and Logs

You can import services from other Datadog telemetry containing the DD_SERVICE tag to populate the Software Catalog. To discover kind:service entities through Datadog infrastructure metrics or logs, navigate to the Software Catalog Import Entries tab.

After importing, entries appear in the Explore tab. Entries may expire unless you add metadata, such as the owner or contacts, by using the API or the GitHub integration.

To remove imported services from the default Explore view, click Clear Previously Imported Services on the Import Entries tab. This removes all services that do not have metadata or do not have APM, Universal Service Monitoring (USM), or Real User Monitoring (RUM) telemetry.

Further reading

Documentation, liens et articles supplémentaires utiles:

Log processing pipelinesDOCUMENTATION