Skip to content
BrainRoad BrainRoad
Documentation Menu

Reading Logs

How to interpret your agent's activity logs and spot issues.

On this page

Log Format

Each line in the Activity feed is a timestamped entry from your agent’s process. Entries include a timestamp prefix followed by the log message. Most lines are plain text describing what the agent is doing at that moment.

Common Log Entries

Here is what typical activity looks like:

  • Message received from WhatsApp — the agent logs the incoming message source and begins processing it.
  • Sending reply — after the agent composes a response, you see it dispatch the message back through the originating channel.
  • Tool use — when the agent browses a website, runs code, or reads a file, each step appears as a separate log entry.
  • Channel connected/disconnected — status changes for WhatsApp, Signal, iMessage, or email connections.
  • Errors — failed API calls, authentication problems, or rate limits appear with error-level markers. These are the entries to watch for.

Spotting Problems

If your agent stops responding to messages, check the Activity feed first. Look for repeated error entries, channel disconnection messages, or the feed going silent entirely. A silent feed usually means the agent process has stopped — try restarting it from the agent detail page.

Authentication errors on a channel mean you may need to re-link that service through the setup wizard.

Activity Tab vs. Console

The Activity tab is a read-only stream of your agent’s standard output. It shows you what the agent is doing. The Console is an interactive shell inside your agent’s container where you can run commands, inspect files, and debug issues directly. Use Activity for monitoring, Console for hands-on troubleshooting.

  • Common Issues — quick fixes for frequently encountered problems