Monitor workbench
Test saved or draft checks without changing monitor health.
The monitor workbench executes saved or draft configurations on a monitoring worker. These runs do not change monitor status, incidents, or uptime.
Use Test configuration in the create/edit drawer to test a draft before saving it, or Run now on an existing monitor to capture an isolated result. Result history on a monitor row opens the workbench with its past ad-hoc results.
Run a test
A run uses the draft exactly as shown, including type-specific assertions, region/agent assignment, address-family mode, retries, proxy, and custom trust. The job is asynchronous: an eligible worker claims it, returns one or more family results, and the workbench updates as results arrive.
Wait for the result cards to appear. Each card shows the observed health, duration, and attempt details. While the job is queued, the workbench shows Waiting for an eligible worker…. During execution, it shows Worker is running the check….
If the job expires, open Agents and compare worker health and assignment with the monitor's region and agent settings. Restore the worker connection or correct the assignment, then select Test configuration for a draft or Run now for a saved monitor. Expired jobs do not count as production failures.
Select Close to dismiss the workbench. An accepted job continues until it completes or expires.
Workbench results are stored separately from scheduled check results. They never update monitor status, confirmation/recovery state, incidents, status-page components, notifications, SLO burn, or ordinary analytics.
Read a result
Each result records source, worker region and ID, address family, total duration, protocol status code, redacted error, attempt/retry counts, and per-attempt timing. Diagnostic-capable workers may add:
- DNS/connect/TLS/TTFB/transfer phase timing;
- assertion expected and actual values;
- redirect hops;
- certificate details;
- a bounded response sample;
- API-sequence step results and extracted variables;
- traceroute details when available.
Sensitive values are excluded from displayed results. Existing-monitor history combines scheduled and workbench observations and annotates nearby deployments and incidents active at check time. Use those annotations as investigation context; a workbench result does not become incident evidence automatically.
Recipes and history
The Templates dialog on the Monitors page offers built-in recipes for common checks: website availability, JSON API health, certificate expiry, DNS resolution, database connection, and scheduled heartbeats. A recipe opens in the normal editor so targets and assertions can be reviewed and tested before creation.
Configuration history offers a side-by-side comparison and an audited restore. See Monitor configuration history.