Skip to main content

How to Handle Leave and Fairness in an On-Call Rotation

GuideWritten by oncall.fyi editorialPublication approved by Burak YApproval recorded 12 September 2026
Sources and verification
Source dates
Oldest source check: 12 September 2026.
Technical verification
Separate technical verification has not been recorded.

Publication approval and technical verification are recorded separately. Automated link checks establish reachability, not accuracy. A checked example verifies only its stated test cases, not the whole article or your production setup.

In short

Manage leave by recording availability before generating the rota, assigning a capable replacement and checking the final schedule at every changed boundary. Measure scheduled hours, nights, weekends and actual incident effort separately. Review the resulting burden with the team over several rotations, and correct imbalance through agreed swaps or staffing changes. Equal numbers of shifts alone do not establish fair on-call work.

Key takeaways

  • Verify the final primary and backup after each leave change.
  • Count overnight disruption separately from scheduled standby.
  • Fairness needs agreed constraints and a review period.

Collect constraints before assigning people

Start with the period you are scheduling and the people qualified for each service. Ask for unavailable dates, agreed working arrangements and existing cover commitments. The rota needs availability, not private explanations for leave. Keep sensitive personnel details in the appropriate private system.

Specify the required primary and backup coverage and who can authorize a temporary staffing change. If two people are unavailable and the remaining engineer cannot safely cover both roles, record an uncovered requirement. A calendar that repeats their name twice does not create independent backup capacity.

Distinguish a temporary replacement from a rotation change

Use an override for a bounded replacement when your scheduling model supports it. Use a planned rotation edit for a lasting change in membership or handoff time. Avoid building months of recurring exceptions that nobody can reconcile with the base rota.

PagerDuty's current shift-based editing documentation (opens in a new tab) distinguishes replacement overrides from additional coverage. Its legacy schedule model (opens in a new tab) uses layers with precedence. Check which model you operate before assuming that a copied exception creates the same final coverage.

Record the original responder, replacement, exact start and end, timezone, affected services and backup. Confirm that the replacement accepts the responsibility and has working access. A calendar invitation alone is not proof of acceptance or readiness.

Measure several kinds of burden

The following is an illustrative review table for one eight-week period. The figures are invented and should not be used to rank employees.

ResponderScheduled / weekend hoursNights with an incidentActive response hours
A168 / 4812
B168 / 4859
C144 / 2423

A and B have the same scheduled allocation, yet B experienced more disrupted nights and active work. C's lower total might reflect agreed availability or leave rather than a scheduling defect. Discuss these dimensions instead of turning them into one opaque fairness score.

Separate primary and backup hours. A backup who repeatedly becomes the main responder is doing work the schedule view may not show. Also record administrative time spent finding replacements; repeated last-minute swaps can reveal an unsustainable rota.

Review leave and fairness together

Choose a review horizon long enough to include several rotations and holiday patterns. Compare observed allocations with agreed availability and qualifications. A useful question is whether the same person repeatedly receives short-notice cover or the most disruptive weekends, and what scheduling rule caused that pattern.

Agree corrections prospectively: rebalance future weekends, recruit another qualified backup, reduce a noisy service's pages, or shorten an overloaded shift. Avoid pressuring someone to take extra duty to compensate for approved leave. Pay and recovery arrangements belong in the team's applicable policy and employment process; a burden worksheet does not determine those entitlements.

Test the changed boundaries

For a leave interval from Friday afternoon through Monday morning, inspect coverage immediately before the start, at the start, inside the interval, immediately before the end and at the end. Check both roles and every service that uses the affected schedule. Include the local timezone and UTC representation in the evidence.

Create a controlled test incident for the temporary primary and withhold acknowledgement once with both responders informed. Confirm that the backup receives it. Then remove a disposable test override and inspect the restored base coverage. Do not edit a real leave arrangement merely to perform a drill.

Use the schedule-change checklist when timezone transitions or several layers are involved. A spot check inside the shift can miss a gap at its edge.

Make the next rota easier to operate

Publish the resulting coverage, the approved replacement process and the contact for unresolved gaps. Review actual burden at the end of the period and retain only the personnel data needed for that purpose. The aim is a rota that supports reliable response and usable off-call time, consistent with the broader Google SRE discussion of on-call work (opens in a new tab). A mathematically balanced calendar is useful evidence, but the team's experienced workload still needs attention.

Did this help?

Your answer helps us improve this guide. We save only the page and your choice for 30 days.

No name, email, or incident details are requested.

Frequently asked

Is an equal number of shifts fair?
It is one input. Compare scheduled hours, weekends, disrupted nights and actual response work alongside agreed availability.
Can the primary also be the backup during leave?
That removes independent coverage. Record the staffing gap and arrange an explicit alternative instead of treating duplicate names as two responders.

Sources

Vendor facts change. Each source below shows the date this page last checked it.

  1. Being on-call Google SRE. Checked 12 September 2026.
  2. Edit shift-based schedules PagerDuty. Checked 12 September 2026.
  3. Schedule basics (legacy) PagerDuty. Checked 12 September 2026.

Related

One practical idea, occasionally

The On-Call Brief: short field notes, templates, and operational lessons.

Follow the field guide

Email subscriptions are not open yet. Read new guides in your feed reader — no email address needed.

Subscribe with RSS