On-call Compensation
On-call is paid back in time off at 1.5x, with recovery first and the remainder added to your PTO. It covers any incident or war-room work beyond reasonable hours, not only the platform rotation, and anyone pulled in is eligible whatever their function. Severity sets how much rest is required, not the multiplier.
On-call is compensated with time off rather than additional pay, so that anyone who works an incident out of hours recovers properly afterwards.
- What counts as an incident. Any urgent work that cannot wait for the next business day and stretches you beyond reasonable hours. This is not only infrastructure or network incidents. A customer incident that pulls you into war-room work, a production regression, a security event, or a launch that goes sideways all count. This kind of work is inevitable in business; when it lands outside normal hours, it is compensated.
- Active incident work is compensated. Time spent working an incident, from the moment you are pulled in until it is resolved or handed off, earns 1.5 hours of time off for every hour worked (1.5x).
- Availability alone is not. Carrying the pager, or being on standby, during hours in which no incident occurs does not accrue time off. The platform rotation is light (approximately one week per month) and pages are expected to be infrequent once alerting is tuned. This will be reviewed if the on-call burden increases.
- The rotation is for platform engineers. The compensation is for everyone. The scheduled pager rotation is staffed by platform engineers, because they hold the infrastructure. Incident and war-room work is not theirs alone. Anyone in the company pulled into an incident beyond reasonable hours is eligible for this compensation, whatever their function: engineering, marketing, sales, product or anyone else. If a customer incident needs marketing on a communication line through the night, or sales standing up with an account, that time is compensated on the same 1.5x terms.
- This covers the shared rotation and ad-hoc incident work. A dedicated primary on-call role, where on-call is the whole job, is salaried for that availability rather than paid per page. See the Site Reliability Engineer role on Jobs at Intempt.
How the earned time is used
- Recovery takes priority. After a night or early-morning page, take recovery time immediately. This is expected rather than requested, and it is drawn from the time earned.
- The remainder is added to your PTO. Any time left after recovery is credited to your PTO balance.
- If recovery requires more time than was earned, take it regardless. Earned time determines how much is credited, not how much rest you are permitted.
What you get, by time of day
The multiplier is always 1.5x. What varies with the time of the page is whether recovery comes first. A page during core sleep hours (00:00 to 06:00) is far more disruptive than an evening one, so a night page requires recovery before anything else, while an evening page converts in full to PTO.
| Time of page | Immediate recovery | Earned time (1.5x) |
|---|---|---|
| Weekday evening (18:00 to 00:00) | Not required; you were awake | Credited to PTO in full |
| Any night (00:00 to 06:00) | Required: start no earlier than 13:00, or take the morning | Recovery is drawn first; the remainder is credited to PTO |
| Weekend daytime | Not required | Credited to PTO in full |
| Weekend night (00:00 to 06:00) | Required: take the morning | Recovery is drawn first; the remainder is credited to PTO |
Severity is reflected in whether recovery is required, not in a larger multiplier.
The rotation
With four platform engineers on the rotation, no one is on duty more than roughly one week per month, and additional members reduce each person's frequency further. If the rotation falls below four, notify People Ops: a thin rotation is a reliability concern before it is a staffing one. This section is the scheduled rotation only. Incident and war-room compensation above applies to anyone pulled in, on or off the rotation.
Recovery is the mandatory element of this policy. Fatigue on a rotation is the primary risk to reliability, so rest takes precedence over the multiplier.