Alerts

Alerts notify the team when something goes wrong, and again on recovery.

Creating an alert

latency alert create \
  --monitor my-monitor \
  --channel slack \
  --on failure,recovery

Channels

Channel	Setup
Email	Verified address
Webhook	Any HTTPS endpoint (Hobby+)
Slack	Incoming webhook URL (Pro+)
Discord	Incoming webhook URL (Pro+)
SMS	Phone number (Pro+)

When alerts fire

Each monitor decides when failures become an incident: a number of consecutive failures opens the incident and a number of successes closes the incident, with multi-region confirmation. Alerts go out when the incident opens and again on recovery, so a blip that clears by itself stays quiet.

What an alert says

Every monitor has a priority (P1 to P5), can alert on warnings and recoveries, can repeat while failing, and can carry a custom message with variables like monitor.name, state and cause. Contact methods can limit delivery to the more urgent priorities. See Notifications; notifications work the same for check and telemetry monitors.

Alerts on telemetry

Telemetry monitors alert on a search count, a field's p95, any ingested metric, or a heartbeat that stops. See Telemetry Monitors and Heartbeat Monitors.