Rotations fail at the boundary. The outgoing person knows the weird flake on payments. The incoming person finds out at 2am. A meeting could fix that, and will be cancelled three weeks in a row.
The async template
Post in #on-call (or your doc) at the handoff time:
- Incoming / outgoing names and timezone
- Open incidents or "none"
- Watch list: deploys in flight, elevated error rates, vendor drama
- Access reminders: anything that broke last week (VPN, vault, cloud role)
- Overrides: swaps already on the calendar
Incoming replies with ack. No ack within a few hours → ping once. Still nothing → escalate to whoever owns the rotation.
Timing
Pick a boring weekday hour overlapping both people when you can. Monday 10:00 local for a weekly rotation is a common pattern. Follow-the-sun teams hand off at region boundaries with the same written form.
Tallwatch stores timezones but does not yet auto-shift handoffs to local clocks, so be explicit in the note if you span regions (rotation guide).
What not to include
- Full architecture tours
- Tone policing about last week's incident
- "All green, nothing to say" deleted as empty; "nothing open, payments still noisy on retries" is better
Make it habit
Calendar reminder for the outgoing person. Checklist in the runbook index. If handoffs skip, pages get worse. Treat a missed handoff like a process bug in the next retro.
You do not need a meeting. You need the next human less blind than you were.
