Opsgenie
InfraSage pages through the Opsgenie Alerts API. High-severity alerts create Opsgenie alerts with a stable alias, your routing rules and schedules take over, and InfraSage closes the alert when the incident resolves.
Configuration
One environment variable on the engine:
OPSGENIE_API_KEY=<an Opsgenie API integration key>
Create the key under Settings → Integrations → API. Unset means the channel doesn't exist. The same severity floor applies as for PagerDuty:
PAGER_MIN_SEVERITY=high # low | medium | high | critical
What gets sent
- message:
[severity] service — message - priority: mapped to Opsgenie's scale (
critical→P1,high→P2,medium→P3) - alias: the alert's
service:severityfingerprint, so repeat fires de-duplicate into the same Opsgenie alert instead of re-notifying - description: the evidence lines
- details: service id, causal incident id, and a dashboard link
Resolution
When the incident resolves in InfraSage, InfraSage closes the alert through its alias and leaves a note saying it auto-resolved. Routing rules, schedules, and escalations stay entirely Opsgenie's.