On-call handover template: what to hand over at every shift change

A good on-call handover takes five minutes and answers four questions: what is broken now, what might break soon, what changed, and who to ask. Copy the template below into your team chat or ticket, fill it in at each shift change, and delete the lines you do not need.

This is a plain-text template, not a tool, and it works with any rota: a spreadsheet, a calendar, or a planner such as Turnkeep. It follows the shape that Google's SRE book recommends for on-call hand-offs: the outgoing person writes down what the incoming person cannot see from dashboards. See the chapter on being on call and the one on managing incidents.

Copy-and-fill template

ON-CALL HANDOVER
Rota: [primary / secondary / support desk]
From: [outgoing name]   To: [incoming name]
Handover time: [YYYY-MM-DD HH:MM, team time zone]

1. OPEN NOW
- [incident or ticket link] - status, what was tried, next step, who owns it
- (none) if nothing is open

2. WATCH LIST
- [alert that flapped, disk filling, certificate expiring, deploy pending]
- What would make it worse, and what to do first

3. WHAT CHANGED THIS SHIFT
- Deploys, config changes, silenced or muted alerts (and when they unmute)
- Temporary workarounds that must be undone

4. KNOWN NOISE
- Alerts that fire but need no action, and why

5. PLANNED
- Maintenance windows, releases, holidays and anyone unreachable
- Swaps already agreed: [who covers whom, which days]

6. WHO TO ASK
- Escalation contact, and a subject expert for each open item

Read back: incoming person confirms in one line that they have the pager, the access and the links.

Why each section is there

• Open now comes first because it is the only part that is urgent in the first ten minutes of a shift.
• Watch list holds things that are not incidents yet. Without it, the incoming person finds out about a filling disk at 3 a.m.
• What changed is the most useful section in practice, because most incidents follow a change. Muted alerts matter most: write down when they unmute.
• Known noise stops a new person from spending an hour on an alert the last person already judged harmless. Keep each line to one reason, and review the list weekly so noise gets fixed instead of inherited.
• Planned includes swaps. A swap that lives only in two people's heads is the usual cause of "I thought you had it".
• Read back catches the cases where the incoming person has not got the pager set up, a VPN login, or a password manager item.

Where to post it

Pick one place and always use it: a pinned thread in the on-call chat channel, a recurring ticket per shift, or a shared document with one section per week. A fixed place matters more than the format. Keep handovers for at least a quarter, so that you can see which watch-list items keep coming back; those are the problems to fix.

Timing and fairness

Hand over at the same local time every shift, for example Monday 09:00 in the team's time zone, and put that time in the calendar feed so nobody has to ask. Handing over at midnight UTC for a team in another zone is a common source of confusion, especially across clock changes. Fixed handover times also make the load easy to count: see how to make an on-call rota fair for weighting weekends and holidays, and the team size calculator to check that your rota is not asking too much of too few people.

Where Turnkeep fits

Turnkeep plans and publishes the rota: who is on, when, with a fairness table, a calendar link and a "who is on now" link you can paste into a runbook or chat bot. It does not store handover notes yet, so keep using the template above wherever your team already talks. It does not page anyone. The generator is free; the paid plan is a flat $6 per month per team.