Skip to content
DT Nova

On-call scheduling software: rotations, standby rules and tools

On-call scheduling software decides who answers when something breaks outside normal hours, and what happens if they do not. For an IT team that means incident alerts and escalation; for a hospital, a maintenance company or a utility it means out-of-hours duty that sits next to the regular rota. This guide covers rotation patterns, how many people a rotation needs, when standby counts as working time in the EU and the US, pay and rest, and how to choose between the established tools and custom software.

What on-call scheduling covers

An on-call schedule has four jobs. It says who is on call for each service or area at each moment, including overrides when someone swaps or falls ill. It reaches that person through the right channel, such as a push notification, a text message or a voice alert. It escalates to a secondary person or a manager when the first person does not acknowledge in time. And it records what happened: when the alert came, who responded, how long the work took and whether the person had their rest afterwards.

Two worlds use the term. In IT operations, on-call scheduling is part of incident response, tied to monitoring and alerting tools. In healthcare, emergency services, building maintenance, utilities and property management, on-call duty is a type of shift with its own pay rate and its own working time rules, and the hard part is fitting it into the regular schedule without breaking rest periods.

Common rotation patterns

PatternHow it worksFits
Weekly rotationOne person on call for a full week, handing over on a fixed dayLow alert volume, small teams
Primary and secondaryA second person takes alerts the first does not acknowledge, and helps on large incidentsServices where a missed alert is costly
Follow the sunTeams in different time zones cover their own daytimeTeams spread across regions
Weekday and weekend splitSeparate rotations for working days and weekends or holidaysFairer spread of the heaviest duty
Night duty from homeOn call outside shifts, with a response time and a recall ruleHospitals, maintenance, utilities

Whatever the pattern, a handover note at the end of each rotation, a clear escalation path and a written rule for swaps prevent most of the friction.

How many people a rotation needs

Google's Site Reliability Engineering book gives one of the few published rules of thumb. Google aims to spend at least 50% of SRE time on engineering work and no more than 25% on call. From that, assuming a primary and a secondary in week-long shifts, the minimum single-site team for 24/7 on-call duty is eight people, each on call one week a month; for teams split across two sites, a reasonable minimum is six per site. The book also sets a load limit: at most two incidents per 12-hour on-call shift, so that each one can be handled properly and followed up (Google SRE book, Being On-Call, read on 11 November 2026).

Those numbers come from one company's engineering practice, not from law, but the logic transfers: count the hours each person spends on call per month, count the alerts per shift, and treat a rising count as a staffing problem, not a personal one.

When standby counts as working time

In the EU, the Court of Justice has ruled several times on standby under the Working Time Directive. In Matzak (C-518/15, 2018), standby at home with a duty to respond within 8 minutes was working time. In two Grand Chamber judgments of 9 March 2021, the Court set out the general test. Standby during which the worker only has to be reachable by telephone and able to return within one hour is not working time in its entirety, unless an overall assessment shows that the constraints affect objectively and very significantly the worker's ability to manage their free time (C-344/19). A firefighter who must reach the town boundary within 20 minutes in uniform with a service vehicle has working time only if the same overall assessment, including the average frequency of interventions, shows such constraints (C-580/19). The time actually spent working during standby always counts. The Directive also requires a minimum daily rest of 11 consecutive hours per 24-hour period.

In the United States, an employee who must remain on call on the employer's premises, or so close that they cannot use the time effectively for their own purposes, is working while on call; an employee who only has to leave word where they can be reached is not (29 CFR 785.17). (All sources read on 11 November 2026.)

For a scheduling system, this means the response time, the expected frequency of alerts and the place the person has to stay are data, not footnotes. They decide whether hours count toward weekly limits, how they are paid and whether the next day's shift starts on time.

Pay, rest and fairness

On-call duty is usually paid in two parts: a standby allowance for being available, and ordinary or overtime pay for the time actually worked, often with a minimum per intervention. A good schedule records both automatically from the alert and response data instead of from claims filled in at the end of the month. It also protects rest: if someone worked through the night, the system should flag the next morning's shift, and the rota should spread nights, weekends and public holidays by a rule everyone can see. Teams that track alert counts per person and per shift spot burnout before it becomes resignations.

Established tools

For IT incident response, the main products include PagerDuty, Jira Service Management, incident.io, Splunk On-Call and the on-call features built into monitoring platforms. Teams on Opsgenie have a deadline: Atlassian ended its sale on 4 June 2025 and support ends on 5 April 2027, with an automated move to Jira Service Management, and Opsgenie data that has not been moved is deleted at shutdown (Atlassian, Opsgenie migration, read on 11 November 2026). For general shift work, rota tools such as Microsoft Teams Shifts or dedicated scheduling products can hold on-call shifts, but most do not escalate alerts.

Questions worth asking any vendor:

  • Can overrides and swaps be made from the phone, and are they checked against rest rules?
  • How does escalation work when nobody acknowledges, and is every step logged?
  • Can standby allowance and time worked be exported to payroll, by your rules?
  • Where is the data hosted, and can you export it in full if you leave?
  • Does it connect to the monitoring, ticketing or building systems that raise the alerts?

When custom on-call scheduling makes sense

For a software team that needs alerts from monitoring, a product is almost always the right choice. Custom software pays off when on-call duty is part of a wider schedule with its own pay rules, qualifications and working time limits. A hospital needs the on-call consultant in the same roster as the day shifts; an ambulance service needs standby counted toward its work period; a property or maintenance company needs the on-call technician to receive the alarm, see the site history and log the job. In those cases the on-call rota belongs in the same system as the shifts and payroll, and we build it as part of employee scheduling software, healthcare scheduling software, EMS scheduling software, police scheduling software or field service management software. For IT operations without a team of your own, see managed IT services. For the wider loop around on-call rotas, from forecasts to time and attendance, see workforce management software.

The short version

What is on-call scheduling software?

It plans who is on call when, notifies that person when something needs attention, escalates to the next person if nobody responds, and records the time spent so it can be paid and checked against rest rules. IT teams use it for incident response; hospitals, maintenance teams, utilities and property managers use it for out-of-hours duty.

How many people do we need for a 24/7 on-call rotation?

It depends on the load and the rules you follow. Google's SRE book derives a minimum of eight engineers for a single-site team with a primary and a secondary on call in week-long shifts, so each person is on call one week a month, and six per site for a team split across two sites. Smaller teams can run 24/7 rotations, but at a cost to rest and retention.

Is on-call time working time?

In the EU, standby time counts as working time in full when the constraints on the worker objectively and very significantly affect their ability to use the free time for their own interests, for example a duty to respond within a few minutes. In the US, on-call time is working time when the employee must stay on or so close to the premises that they cannot use the time effectively for their own purposes. Collective agreements and national law add their own rules.

What happens to Opsgenie?

Atlassian ended the sale of Opsgenie on 4 June 2025 and support ends on 5 April 2027. Atlassian offers an automated move to Jira Service Management, and data not moved by then is deleted, so teams on Opsgenie need to plan a migration or choose another tool before that date.

Should we build our own on-call scheduling software?

Usually not for IT incident response, where mature products exist. Custom software makes sense when on-call duty is part of a wider schedule with pay rules, qualifications and working time limits, for example in healthcare, emergency services, field maintenance or property management, and the on-call rota has to sit in the same system as the shifts and payroll.

Sources