OpenTelemetry Implementation
OpenTelemetry provides a standardized way to collect traces, metrics, and logs from applications. When used with Atatus, OpenTelemetry acts as the instrumentation layer, while Atatus becomes the backend that ingests, correlates, and visualizes telemetry data. This approach is especially useful for teams that want vendor-neutral instrumentation without rewriting observability logic later.
Setup and Configuration
Decide which telemetry signals to collect such as traces, metrics, or logs.
Install the OpenTelemetry SDK and API packages for your application runtime.
Initialize OpenTelemetry in the application entry point before the framework loads.
Define service name, environment, and version for correct grouping in Atatus.
Configure the OTLP exporter to send telemetry data to the Atatus ingestion endpoint.
Enable only the required telemetry pipelines to avoid unnecessary overhead.
Configure sampling rules to control the volume of trace data.
Integration Points
Libraries and frameworks are instrumented automatically to emit spans for incoming requests, outgoing HTTP calls, database queries, and background jobs.
Custom spans can be added around business-critical operations to capture timing and context that automatic instrumentation cannot infer.
Application-specific counters, gauges, and histograms can be recorded using OpenTelemetry meters and exported directly to Atatus.
Logs emitted with trace and span context can be linked to distributed traces inside Atatus for faster root-cause analysis.
Telemetry can be sent directly to Atatus or routed through an OpenTelemetry Collector for preprocessing, filtering, or multi-destination routing.
Testing and Validation
Run the application in a development environment with OpenTelemetry enabled.
Send requests, execute background jobs, and perform database operations to generate telemetry.
Confirm that distributed traces appear in Atatus with correct service names and operation breakdowns.
Ensure parent-child span relationships are preserved across services.
Confirm that environment, version, and custom attributes are consistently attached to telemetry.
Verify that sampling rules behave as expected under normal and high-traffic scenarios.
Key Takeaways
- OpenTelemetry separates instrumentation from backend, giving flexibility without losing observability depth.
- Early SDK initialization is essential for complete and accurate telemetry capture.
- Automatic instrumentation provides fast coverage, while manual spans fill visibility gaps.
- Consistent service metadata is critical for usable traces and metrics inside Atatus.
- Testing in non-production environments ensures telemetry accuracy before scaling.