OrbitWard/docs/alerting-design.md

Alerting Design

Alerting is built around alert rules, incidents, notification policies, and notification history.

Alert Rules

An alert rule turns monitor status or metric data into an incident. Initial rule behavior should support:

• Failure thresholds
• Recovery notifications
• Cooldown
• Severity
• Acknowledge
• Silence

Incidents

Incidents represent active or historical alert events. They include opened time, resolved time, current status, severity, related asset, related monitor, related alert rule, notification history, acknowledgement, and silence state.

Notifications

Initial channels:

• Email / SMTP
• Mattermost incoming webhook
• Zoom Team Chat incoming webhook
• Generic webhook

Alert messages should be human-readable and include asset, check, status, duration, timestamps, and a link back to InfraPulse.