Alert rules

Server metric alert rules: metrics, comparisons, durations and recovery notifications.

Alert rules watch the metrics your agents report. A rule fires when its condition holds for notify_after minutes, opening an incident and notifying its recipients.

Fields

Field	Description
`server_ids`	Servers to watch. Empty array = all servers, including ones added later.
`metric`	See below.
`compare`	`HIGHER`, `LOWER`, `EQUAL`, `DOESNT_EQUAL`
`value`	Threshold.
`notify_after`	Minutes the condition must hold before alerting.
`notify_on_close`	Send a recovery alert when the condition clears.
`channel_ids`	Recipients to notify.

Metrics

Metric	Meaning	Unit
`down`	Server stopped reporting (no compare/value needed)	—
`CPU_overall`	CPU busy	%
`MEM_used`	Memory used	%
`MEM_free`	Memory free	%
`CPU_temp`	CPU temperature	°C
`CPU_loadavg`	1-minute load average	—
`DISK_used`	Highest usage across mounts	%

Example

curl -X POST https://api.statance.com/v1/alerts \
  -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" \
  -d '{"server_ids":[],"metric":"CPU_overall","compare":"HIGHER","value":90,
       "notify_after":10,"notify_on_close":true,"channel_ids":["ch_..."]}'

Next: Notifications →