Skip to content
Coder Docs

Logs

Learn how Coder services write logs to standard error and how to collect them.

On this page

All Coder services log to standard error, which can be critical for identifying errors and monitoring Coder's deployment health. Like any service, logs can be captured via Splunk, Datadog, Grafana Loki, or other ingestion tools.

coderd Logs

By default, coderd, the process that runs the control plane, exports human-readable logs to standard error. You can access these logs via kubectl logs deployment/coder -n <coder-namespace> on Kubernetes or journalctl -u coder if you deployed Coder on a host machine/VM.

Note

To disable human-readable logging, set --log-human (or CODER_LOGGING_HUMAN) to /dev/null. If you set no JSON or Stackdriver log location, Coder then writes no logs. Do not use an empty value to disable logging. If every log location is empty and --trace (CODER_TRACE_ENABLE) is off, the server exits with no loggers provided, use /dev/null to disable logging.

Events such as server errors, audit logs, user activities, and single sign-on (SSO) & OpenID Connect logs are all captured in the coderd logs.

provisionerd Logs

Logs for external provisioners are structured and configured similarly to coderd logs. Use these logs to troubleshoot and monitor the Terraform operations behind workspaces and templates.

Workspace Logs

The Coder agent inside workspaces provides useful logs around workspace-to-server and client-to-workspace connections. For Kubernetes workspaces, these are typically the pod logs as the agent runs via the container entrypoint.

Agent logs are also stored in the workspace filesystem by default:

  • macOS/Linux: /tmp/coder-agent.log
  • Windows: Refer to the template code (for example, azure-windows) to see where logs are stored.

Note

The agent rotates its log file once it reaches 5 MB and keeps up to 19 rotated files. For details, refer to Trace connections.

Startup script logs are also stored in the temporary directory of macOS and Linux workspaces.

Kubernetes Event Logs

Sometimes, a workspace might take a while to start or even fail to start due to underlying events on the Kubernetes cluster such as a node being out of resources or a missing image. You can install coder-logstream-kube to stream Kubernetes events to the Coder UI.

Kubernetes logs in Coder dashboard