Skip to main content

Viewing Logs

Logs give you real-time visibility into what your OpenClaw instance is doing. They are useful for confirming that channels are connected, monitoring agent activity, and diagnosing issues.

Accessing Logs​

From the Dashboard​

  1. Go to your OpenClaw Dashboard.
  2. Navigate to Instance.
  3. Click the Logs tab.

The log viewer shows the most recent log entries and updates in real time.

From the CLI​

If you have the CLI installed, you can stream logs directly in your terminal:

openclaw logs --follow

Press Ctrl+C to stop streaming.

What Logs Show​

Logs contain several categories of events:

Gateway Startup​

When your instance starts, you will see entries like:

Gateway starting on port 18789...
Loading plugins...
telegram: loaded
discord: loaded
Gateway ready.

This confirms the gateway is running and which channel plugins were loaded.

Channel Connections​

When a channel connects or disconnects:

telegram: connected as @my_ai_bot
discord: connected to 3 guilds
slack: socket mode connected

Agent Messages​

When your agent processes a message:

[telegram] Received message from user 12345678
[agent] Processing turn...
[telegram] Sent reply to user 12345678

Errors​

Errors are clearly marked and include context about what went wrong:

ERROR: anthropic API returned 401 - Invalid API key
ERROR: telegram connection lost, reconnecting...

Tips for Debugging​

Channel Not Connecting​

Look for error messages immediately after the plugin loads. Common patterns:

  • 401 Unauthorized -- your bot token is incorrect or expired
  • Connection refused -- a network issue, usually resolves on restart
  • Missing scope -- the bot does not have required permissions on the platform

Agent Not Responding​

Check for errors related to your AI provider:

  • 401 or 403 errors mean your API key is invalid or lacks permissions
  • 429 Too Many Requests means you have hit a rate limit
  • insufficient_quota means your AI provider account has no remaining credits

Instance Restarting​

If your instance keeps restarting, look for crash logs right before the restart:

  • FATAL or unhandled exception entries indicate a bug -- contact support
  • OOMKilled means the instance ran out of memory -- consider upgrading your plan
tip

When contacting support, include the relevant log entries. This helps us diagnose your issue much faster.