OpenTelemetry (OTEL) collection provides you with distributed tracing, metrics, and logs across your Daytona sandbox operations and SDK calls.
Daytona now covers both control-plane and runtime observability: SDK operations, API requests, and sandbox activity can all be exported through OpenTelemetry. It provides your Daytona organization a telemetry path for both platform operations and workload behavior in a format that works with standard OTLP-compatible backends. This makes it practical for capacity planning, quota reviews, and identifying saturation before it affects workloads.
Observability for sandboxes
As agents and automated pipelines run code inside sandboxes at scale, understanding what is happening inside them requires direct visibility. OpenTelemetry covers two distinct layers: telemetry collected from the sandbox runtime itself, and traces generated by the Daytona SDK as it interacts with the API.
Sandbox telemetry. Collect CPU, memory, and filesystem metrics alongside application logs and traces directly from running sandboxes. View them in the Dashboard under the Logs, Traces, and Metrics tabs on any sandbox.
SDK tracing. Trace all SDK operations including sandbox creation, start, stop, deletion, file system access, and code execution. Enable it with a single flag when initializing the Daytona client.
Organization-level metrics. Get a high-level view of CPU, memory, and storage consumption across your entire organization, exported automatically every 60 seconds to your OTLP endpoint.
Bring your own backend. Route telemetry to New Relic, Grafana Cloud, Jaeger, or any other OTLP-compatible collector. Dashboard examples for New Relic and Grafana are available in the Daytona GitHub repository.
Get started
OpenTelemetry is available for all Daytona organizations. Read the OpenTelemetry Collection documentation to get started.
