Managing Monitors
Creating a monitor​
- Go to Monitors in the dashboard
- Click New Monitor
- Fill in the form:
- Name — a descriptive name for the job
- Expected Schedule — the cron expression, e.g.
0 9 * * * - Grace Period — seconds to wait before raising a missed-run alert
- Timezone — the timezone for your cron expression (default: UTC)
- Alert Channels — which channels to notify on failures. Click + New channel to create one inline without leaving the page.
- Tags — optional labels for organising your monitors
- Alert Rules — optional conditions that fire a silent failure alert based on job output
- Click Create Monitor
After creation, copy the Monitor ID from the detail page — you'll need it in your SDK configuration.
Reading the monitor detail page​
The monitor detail page shows:
- Status badge — current health at a glance
- Success rate — percentage of successful runs in the last 30 days
- Average duration — mean run time across successful runs
- p95 duration — 95th percentile run time (useful for setting grace periods)
- Duration trend chart — run durations plotted over 30 days; anomaly runs are highlighted
- Run history — paginated list of all runs, filterable by status
Click any run to open the run detail panel, which shows:
- Start time, end time, and duration
- Error message (for failed runs)
- Attached log output (for failed runs with a log — expandable/collapsible)
- Whether the run was a duration anomaly
Alert rules​
You can configure up to 5 alert rules per monitor. Each rule watches a numeric field in the run's metadata and fires a silent_failure alert when the condition is true.
To add a rule:
- Open the monitor form (create or edit)
- Scroll to Alert Rules
- Enter a field name, choose an operator (
lt,gt,eq,ne), and set a threshold - Click + Add Rule
Rules are evaluated on every success ping that includes the relevant field. See Alert Rules for full details.
Pausing and resuming​
Click Pause on the monitor detail page to stop alerts from firing without deleting the monitor. Click Resume to re-enable it.
Use this during planned maintenance windows or deployments.
Editing a monitor​
Click Edit to update the name, schedule, grace period, timezone, alert channels, tags, or alert rules.
Deleting a monitor​
Click Delete on the monitor detail page. This permanently deletes the monitor and all its run history. This action cannot be undone.
Deleting a monitor does not disable the SDK integration in your job code. If your job keeps sending pings to the deleted monitor ID, they will return a 404 error — silently handled in default silent mode.