On-Call and After-Hours Work
What you'll learn: How to manage on-call rotations and incident response fairly while protecting team health and sustaining high performance.
The Reality of Production Ownership
When you embrace production ownership, someone must be available when systems fail at 2 AM. On-call work is necessary—but it can destroy sustainable pace if managed poorly. The challenge is balancing business needs for reliability with engineers' need for recovery and personal time.
Fair On-Call Structures
Rotation Design: Distribute on-call duties fairly across the team. A typical rotation lasts one week, with clear handoff ceremonies. No one should be perpetually on-call, and junior engineers need support through shadow rotations before taking primary duty.
Compensation: On-call work deserves compensation—whether financial bonuses, additional PTO, or time-in-lieu after incidents. If an engineer spends Saturday night troubleshooting production, they shouldn't be expected at their desk Monday morning.
Incident Thresholds: Not everything warrants a 3 AM page. Define clear severity levels with appropriate response expectations. P0 incidents get immediate pages; P3 issues can wait until business hours.
Recovery and Prevention
Follow-the-sun: Global teams can pass on-call responsibilities across time zones, reducing after-hours burden.
Post-incident recovery: After major incidents, affected engineers need explicit time off—ideally the next business day. Treat this like a workplace injury requiring recovery.
Reduce the pain: Track what causes pages and ruthlessly fix repeat offenders. Good on-call management makes being on-call boring through better systems, not heroic firefighting.
Key Takeaway: On-call work sustains production reliability but requires fair rotation, adequate compensation, mandatory recovery time, and continuous investment in reducing incident frequency—otherwise you burn out the people keeping your systems running.