Keep the workflow small but explicit
A small team can often survive with chat messages and a shared calendar until alerts become frequent, ownership becomes unclear or handoffs start to fail. IncidentRelay gives the team a defined path from alert intake to responder action without forcing a heavy enterprise process.
The basic setup is simple: create a team, create a rotation, attach channels, add one or more routes and send alerts from the monitoring system.
What the team gets
- One current on-call owner from the rotation schedule.
- Route matchers so alerts go to the right team instead of a noisy shared channel.
- Acknowledge, resolve and silence actions for incident hygiene.
- Escalation policies when the first responder does not react.
- Incident history that remains visible after the chat scrolls away.
A practical first workflow
Start with one production route and one notification channel. Match only critical alerts first, verify the selected responder, then add lower-severity notifications when the team is confident in the routing.
| Step | IncidentRelay object |
|---|---|
| Who owns alerts? | Team and rotation |
| Which alerts enter? | Route source and matchers |
| Where do messages go? | Notification channels |
| What happens if nobody responds? | Escalation policy |
Why self-hosting matters here
Small teams often care about cost and control. A self-hosted setup keeps alert data, user data and operational history inside the environment the team already manages.