Use this page at the end of every on-call shift or rotation, and when starting one. A good handoff takes ten minutes and saves the next person an hour of rediscovery at 3 a.m.
Handoff template
On-call handoff: <outgoing> -> <incoming>, <YYYY-MM-DD>
Open incidents:
- <link> <one line: status, next step, owner>
Watch list (fragile, not broken):
- <service>: <why, what to look at, link to dashboard>
Alerts that fired this shift:
- <monitor>: <count>, <actionable? tuned? ticket link>
Scheduled this week:
- <date UTC> <change or release, link, rollback ready?>
Silences and downtimes in place:
- <scope>, ends <time UTC>, reason <link>
Pending on others:
- <ticket>: waiting on <role>, since <date>
Notes:
- <anything the next person must know>
Before you hand off
- Every open incident has a written status and a named owner.
- Every silence or downtime you created has an end time, or is removed.
- Noisy alerts from this shift have tickets (see alert hygiene).
- Postmortem tickets opened for any SEV1 or SEV2 (postmortems).
When you start
- Read the handoff and ask questions while the outgoing person is still around.
- Confirm paging reaches you: send a test page.
- Check access: VPN, SSO, read-only and break-glass roles, deploy rights for rollback.
- Open the main dashboards once, so they load fast later.
How an agent helps
An agent can draft most of the handoff from systems of record: incidents from the incident tool, alerts from the monitoring API, changes from the change calendar, silences from the downtime API, and pending tickets from a saved query. The human edits and sends it. See the daily and weekly report jobs in the automation blueprint.
Pitfalls
- Handoffs only in chat. Put them in one searchable place.
- "Nothing happened" handoffs that skip the silences in place. A forgotten silence is how the next outage goes unnoticed.