Incidents and on-call

Escalation that walks push, SMS, phone call until a human acknowledges, and an incident record that keeps the whole story in one place.

The ladder

Push first (native iOS app and Apple Watch, currently in TestFlight), then SMS, then a phone call, with stage timings per schedule and calls capped at 2 per incident. Numbers must be verified by one-time code before they can ever be paged. SMS and calls from Sentinel. Alerting is platform-neutral: every phone is reachable, the iOS app is a bonus.

Acknowledgment is ownership

Acknowledging stops the escalation, quiets every other device, and addresses reminders to the owner instead of the whole team. The acknowledgment lands on the timeline with its timestamp, next to region votes, notifications, and recovery.

On-call schedules

Rotations with escalation stages and a coverage-gap warning when nobody is on duty. From Sentinel, like incident command in general.

Manual incidents and post-mortems

Open an incident by hand, with or without a bound monitor; curate it; publish it explicitly to the status pages you choose, never implicitly. Post-mortems attach to the incident with impact, root cause, lessons, and action items with owners, moving from draft to published deliberately. Our own post-mortems policy applies the same register to us.

Also reachable

Signed webhooks (HMAC in X-Datargo-Signature, secret rotation, test fire) and Slack via incoming webhook fire on incident open and resolve. The full channel table with the honest gaps is on integrations.

Read the docs, or migrate off Opsgenie before April 2027.