Skip to content

Operations & Troubleshooting ​

Logging ​

ApeAdmin-Gin uses Zap for structured JSON logs. Every request is assigned a unique ID by the request_id middleware, threading through the entire call chain. Operation logs are written through the async auditqueue and never block the main request.

MCP calls are audited automatically, recording request params, response results, latency and caller info — viewable in the admin panel.

When troubleshooting, check in order: startup logs, plugin status, the log entries for the RequestID, and the API responses in the browser Network tab.

Common Issues ​

Check that the menu component resolves to a .vue file of the same name under frontend/src/views (case-sensitive). Then sign out and reload the user menu.

Plugin APIs Missing ​

Check whether the plugin Register() ran successfully and look at the plugin-loading stage in the startup logs. For L2 ZIP plugins, confirm the manifest format is correct. If needed, disable and re-enable.

AI Chat Fails ​

Confirm the AI model API key and network access are configured in the admin panel; check the MCP gateway timeout (default 30s) and concurrency limit (default 10). Inspect the MCP audit logs for error details.

Assets 404 ​

Confirm the plugin is enabled and check the static file directory config (app.spa_dir) and route registration.

Production Checklist ​

  • [ ] Change jwt.secret; otherwise the key circuit-breaker refuses to start (it detects the default secret when debug=false).
  • [ ] Adjust the database pool max_open_conns to match actual load.
  • [ ] Confirm the API rate limit (300/IP/min, burst 50) fits your needs.
  • [ ] Configure CORS Origins for your production domain.
  • [ ] Verify the graceful shutdown flow: plugin unload → event notification → log drain → DB disconnect.
  • [ ] Configure backup and access control for the uploads/ directory.

ApeHub Plugin Ecosystem