- Dockerfile (uv-based) + docker-compose.yml (3 services) - CLI tool 'da' with commands: auth, sync, query, status, admin, diagnose, skills - Scheduler sidecar service (replaces systemd timers) - pyproject.toml for uv distribution - Built-in skills (setup, troubleshoot) for AI agents - 17 CLI tests, 75 total tests passing
1.1 KiB
1.1 KiB
Troubleshoot — Diagnostic procedures
Quick Check
da diagnose --json
Common Issues
Data not updating
da diagnose --component data— check data freshnessda server logs scheduler --since 1h— check scheduler logs- Verify data source credentials:
da admin test-connection
Cannot login
- Check server is running:
curl http://server:8000/api/health - Check user exists:
da admin list-users(from admin account) - Re-generate token:
da login --email your@email.com
DuckDB errors locally
- Re-sync:
da sync(rebuilds views) - Check disk space:
du -sh user/duckdb/ - Delete and re-create:
rm user/duckdb/analytics.duckdb && da sync
Server unresponsive
docker compose ps— check container statusdocker compose logs app --tail 50— check app logsdocker compose restart app— restart app
Escalation
If automated diagnostics don't help:
- Collect full diagnostic:
da diagnose --json > /tmp/diag.json - Collect server logs:
docker compose logs > /tmp/logs.txt - Share both files with admin