snowclaw logs

Fetch container logs from the deployed SPCS service.

Usage

snowclaw logs
snowclaw logs -n 500
snowclaw logs --container cortex-proxy
snowclaw logs --instance 0

Flags

FlagMeaning
-nNumber of log lines to fetch (default: 100).
--containerWhich container to read from: openclaw (default) or cortex-proxy.
--instanceInstance index if the service is scaled to multiple. Defaults to 0.

What to look for

  • OpenClaw gateway — startup banner, WebSocket connections, plugin load errors, request logs.
  • Cortex proxy — inbound requests on /v1/chat/completions and /v1/messages, secret-masking activity, cache-stat lines (Cache stats model=... created=N read=N), any 429 retries.

If the service is in FAILED state, start here. Entrypoint lockdown failures (permission errors on openclaw.json, secrets.json, or the credentials/ dir) will surface in the first 10–20 lines.