Logs Missing

Symptoms

• logs panel is empty for a service you expect to be noisy
• get_logs returns fewer entries than expected
• historical logs seem to disappear after reconnect or restart

Likely causes

• the service is not running or selected scope is wrong
• response was truncated by effective log limit
• you are relying only on live log events and missed some events
• server restarted and in-memory history reset
• disk projection (logs.dir) is not enabled

Resolution

1. Confirm the service status first (running or ready).
2. Request logs without forcing a small client limit:

{
  "type": "command",
  "id": "logs-1",
  "name": "get_logs",
  "payload": {
    "service": "api"
  }
}

3. Check truncated and effective_limit in the result. effective_limit reflects the applied server policy. If the request has no explicit limit, the server applies matching logView override rules, then logs.bufferSize default (5000).
4. For continuity, track last seq and replay with after_seq.
5. If you need session files, set logs.dir and read projected JSONL files.

Expectations to set

• Live log events are best-effort.
• get_logs is the reliable recovery path.
• Without disk projection, history is tied to the current server session.

See Protocol for get_logs command semantics and Configuration for logs.dir and logView fields.