Logging can help you understand what's happening under the hood of your Coder deployment and is useful for debugging and monitoring the health of your cluster.

Accessing logs

You can access your logs at any time by running:

kubectl -n coder logs <podname>

Exporting logs

The following sections show how you can change your Helm chart values to export logs.

See our guide to updating your Helm chart if you're unfamiliar with updating a Helm chart.

Please note that:

• Setting either the /dev/stdout or /dev/stderr value to an empty string to disable.

• You can use /dev/stdout and /dev/stderr interchangeably, since both write to the pod's standard output and error directories.

• Coder supports writing logs to multiple output targets.

Human-readable logs

This is the default value that's set in the Helm chart:

logging:
  human: /dev/stderr

When set, logs will be sent to the /dev/stderr file path and formatted for human readability.

JSON-formatted logs

You can get JSON-formatted logs by setting the json value:

logging:
  json: /dev/stderr

Sending logs to Google Stackdriver

If you deployed your Kubernetes cluster to Google Cloud, you can send logs to Stackdriver:

logging:
  stackdriver: /dev/stderr

Sending logs to Splunk

Coder can send logs directly to Splunk. Splunk uses the HTTP Event Collector (HEC) to receive data and application logs. See Splunk's docs for information on configuring an HEC.

Once you've configured an HEC, you'll need to update your Helm chart with your HTTP (HEC) endpoint and your HEC collector token.

To provide your HTTP (HEC) endpoint:

logging:
  splunk:
    url: ""

To provide your HEC collector token:

logging:
  splunk:
    token: ""

Optionally, you can specify the Splunk channel that you'd like associated with your messages. Channels allow logs to be segmented by client, preventing Coder application logs from affecting other client logs in your Splunk deployment.

logging:
  splunk:
    channel: ""