Observability
Logs panel
The Logs button in the left sidebar opens a resizable bottom panel. It is accessible from any page — Data Explorer, Streams, or connection views.
The panel has two sub-tabs:
- System Logs — real-time and historical logs from the SSE stream, covering stream progress, errors, and diagnostics
- SQL Console — SQL query execution logs with timing and result details
Filtering
- Level filter — narrow logs by severity: errors & warnings, progress & stats, or info
- Search — free-text search across log messages
- Per-stream tabs — each stream gets its own tab so you can isolate its output from other activity
Logs can also be consumed programmatically via the SSE endpoint /logs/stream (see Logs & Monitoring API and SQL Logs API).
Live stream monitoring
Open any running stream and switch to the Monitor tab. It shows:
- Stage and status — current lifecycle state (starting, running, finished, failed, stopped)
- Source reader stats — produced count, data size, transfer rate, elapsed time, failed events
- Target writer stats — consumed count, data size, average rate, elapsed time, active node count
- Per-table progress — row count, data size, status (running / completed / failed / stopped / paused), and progress bar for non-CDC streams
- Upload activity — file staging progress when the target uses S3 or file-based delivery
Warnings and alerts
The Monitor tab surfaces inline alerts when issues arise:
- Log transport degradation warnings
- Active stream blocking alerts (when another stream is already running)
- Evaluation limit warnings (approaching or exceeding trial quotas)
Run history
Open a stream and switch to the History tab to review past executions. Each run shows:
- date and time
- duration
- final status (finished, failed, stopped) — color-coded
- data size transferred
- row counts (inserted and skipped)
Actions per run:
- View Logs — opens the logs panel filtered to that specific run
- Delete — removes a single run from history
- Clear all — removes all history entries for the stream
Use history to compare runs, spot regressions, or verify repeated workflows.