How to set up an on-call rotation for a small team
Published on: November 17, 2026, by: João Alves
Most on-call advice is written for large companies with follow-the-sun teams and dedicated SREs. A team of three to eight has a different problem: there are few enough of you that every decision about the rotation shows up in each person's year. This guide covers the decisions that matter, in the order to make them.
1. Decide what on-call means for your team
Write down three things before you set up any schedule:
- Hours. Is it business hours only, or nights and weekends too?
- Response. How quickly should the person on call react, and to what? "Within 15 minutes, to anything in #alerts" is a very different job from "the same day, to questions from support".
- Scope. Does on-call include triaging bug reports and questions, or only production problems?
If the honest answer is business hours, same-day response and mostly questions, you are running a support rotation, not an incident one. That is good news, because it needs far less tooling.
2. Do the arithmetic
With weekly shifts, each person is on call 52 divided by the team size weeks a year. Six people means about nine weeks each. Four people means thirteen weeks, a quarter of the year. Three means more than seventeen.
Below four people, think hard about what you cover:
- Limit on-call to business hours and accept slower response at night.
- Use daily or weekday shifts, so the load is spread more finely.
- Join a rotation with a neighbouring team, so the pool is bigger.
3. Pick the shift length and the handover
Weekly shifts are the usual choice: long enough to build context on what is breaking, short enough that a bad week ends. Hand over on the same day and at the same time every week, during working hours, so the outgoing person can actually talk to the incoming one.
Many teams hand over mid-week rather than on Monday morning. Nobody starts their shift with a weekend's worth of surprises, and the handover does not collide with the Monday meetings.
4. Name a backup
The person on call will sometimes be on a plane or asleep through their phone. A named backup for each shift removes the "who do I try next?" scramble. A simple pattern is that next week's person is this week's backup, so they start their own week already knowing what is going on.
5. Make swaps cheap and visible
Swaps will happen. The rule that matters is that a swap is written down where the schedule lives, not agreed in a DM and forgotten. Swap a single shift rather than reshuffling the order, so everyone else keeps their dates. Book holidays into the schedule as soon as they are known.
6. Agree on the deal up front
Being on call has a cost, and a small team feels it. Agree before the first shift what the person on call gets: time off after a rough night, a lighter sprint load that week, or extra pay, depending on what your company allows. A rotation that feels unfair gets quietly avoided.
7. Write a one-paragraph handover
At the end of each shift, the outgoing person posts what happened, what is still open and what to watch. It takes a minute and saves the next person an hour of reading alerts.
8. Review it after a month
Look at what actually came in. If nothing paged anyone in four weeks, scale back the hours. If one person got every bad week, check whether that was chance or a pattern in the order. Adjust the rotation, not the people.
Tooling
If people need to be paged, you need a paging tool for that part. For the schedule itself, a rotation with reminders and a shared link is enough, and the on-call schedule generator has example rotations you can copy. If you are weighing up the big platforms, see how RotaHog compares with PagerDuty and Opsgenie.