There are six of us at Tern. We work in Leeds, and we need someone to respond when our own service has a problem outside office hours. A rota helps, but the schedule alone cannot tell you whether the next person has enough context to take over.
Maren Okafor leads engineering and Tobias Lind leads product. Both take part in incident reviews. We assign paging to people who can investigate and make the necessary changes, then make sure a backup can help with the parts they know less well. Being a founder does not make someone familiar with every recent deployment.
Handover has a short written record
Before a handover, the outgoing person writes down open incidents, temporary changes and anything likely to page again. They include links to the incident timeline and the last public update. If a workaround has an expiry time, that time goes in the note.
The incoming person checks their contact details and the next scheduled handover. We send a test alert after changing a phone number or routing rule. A name in a calendar is not enough evidence that the alert will reach the person carrying the phone.
We also confirm the backup. On Business, an escalation policy can notify that person if the primary contact has not acknowledged within the configured delay. The backup needs to know they are on the schedule before the first incident. Holidays and evenings away are easier to cover when we arrange swaps early.
Quiet hours protect routine time
We hold routine reports during quiet hours. Incident paging still follows the rota, so a report arriving in the morning does not mean a failure was held overnight. Keeping those two types of message separate makes the settings easier to explain to a new colleague.
After an overnight incident, the person who responded gets time back during the next working day. We move meetings where we can and give someone else the first investigation task. A six-person team cannot treat sleep as an unlimited resource.
Our weekly review asks whether each page required action, whether the contact had the information they needed and whether the backup was reached at the right time. A repeated false alarm becomes a task to fix the check or the service. We avoid leaving it as a permanent tax on whoever is on call.
The rota is a working agreement that needs maintenance. Clear handovers, tested contact routes and room to recover after an incident make it possible for a small team to keep that agreement. We built those habits before putting their settings into Tern.
