Own the operational loop
DevOps teams usually sit between infrastructure, applications and delivery pipelines. IncidentRelay is built for that operating model: alerts enter from monitoring systems, route by labels or source, resolve to a team and notify the current on-call engineer.
Useful building blocks
- Rotations with layers, membership periods and temporary overrides.
- Routes for Alertmanager, Zabbix, Sentry, LibreNMS and generic webhooks.
- Notification channels for chat, email, browser push, webhooks and voice providers.
- Silences and maintenance windows for planned work.
- OpenAPI endpoints for automation and internal tools.
Routing by labels
Routes can match normalized alert labels such as team, service, severity, namespace or alertname. This lets a DevOps team split production, staging, platform and application alerts without creating a single noisy inbox.
{
"labels": {
"team": "platform",
"env": "prod",
"severity": "critical"
}
}Where it fits
IncidentRelay is a good fit when the team wants a transparent, inspectable on-call system close to the monitoring stack and prefers running core operational tooling itself.