Create operational alerts

Create a rule

Open Monitor → Alerting → Create alert rule. Configure a name, metric, threshold, window, minimum samples and optional agent filter. Set a notification cooldown so the same sustained condition does not repeatedly interrupt your team.

Metrics in the current dashboard cover operational conditions such as latency, failures and concurrency. Choose a threshold appropriate to your actual traffic. A single test call should not carry the same operational weight as a sustained increase across many calls.

Add a destination

Use notification settings to configure an authorized email, Slack or webhook destination where available. A destination being typed into a rule does not prove its delivery credentials are valid. Use a controlled alert condition, inspect its delivery result and confirm the intended person or channel receives it.

A webhook must be an approved public HTTPS endpoint. Do not put a password into an alert label or URL query string. Keep receiving credentials in the designated connection fields.

Verify the whole path

1. Save a rule with a small, controlled test condition.
2. Confirm the condition is evaluated after enough samples exist.
3. Check the alert event and delivery result.
4. Acknowledge it using your team's operating procedure.
5. Restore a practical production threshold.

Alerts depend on job processing, notification configuration and source metrics. They are not an emergency dispatch system. Assign a staff fallback and review failed notification deliveries instead of assuming no notification means no incident.