The Five-Minute On-Call Handoff Pattern

Most on-call handoffs are ‘everything was quiet, you’re up.’ Five minutes of structure prevents the next incident from rediscovering yesterday’s context.

Why handoffs lose context

Outgoing on-call carries a week of live context: which alerts are flaky, which mitigations are still in place, which deploys are mid-flight. Incoming carries none. Without transfer, every recurring incident gets re-investigated as if new.

• Context evaporates fast. Friday afternoon’s "I muted that alert because it’s a known false positive" is gone by Monday morning.
• Recurring incidents repeat. Same alert, same root cause, same investigation; the team rediscovers what someone already learned.
• Mitigations get forgotten. A temporary autoscaling cap, a rate limit bump, a feature flag flip; nobody removes them because nobody knows.
• The fix. Five minutes of structured transfer prevents days of repeated work; the ROI is enormous and the cost is trivial.

Five-minute template

• 1. Active incidents (1 min).
• 2. Recent alerts that fired (last 24h) (1 min).
• 3. Open mitigations / temporary changes (1 min).
• 4. Heads-up on planned changes (1 min).
• 5. Questions (1 min).

Sync vs async

The default mode is async handoff via a written doc; sync handoff via a 5-minute call kicks in when context is unusual. Both modes work; pick deliberately based on what happened during the shift.

• Sync handoff (5 min call). High-context periods: active incident, mid-flight migration, recent on-call change of plan.
• Async handoff (written doc). Quiet periods: no active issues, recent shift mirrored last; reading the doc is enough.
• Default. Written doc plus optional 5-min call if any item is unusual; outgoing on-call decides whether to call.
• The signal. If outgoing reaches for the call, incoming pays attention; the request itself communicates urgency.

The handoff doc

The handoff doc lives in one durable place. Slack DMs do not count: scrollback evaporates, and the doc must survive team turnover and tool migrations.

• Durable location. Confluence, Notion, a markdown file in the runbooks repo; not Slack, not email, not memory.
• Linked from rotation system. PagerDuty / Opsgenie schedule entry links the doc; one click from "who is on call" to "what they handed me."
• Append-only history. Each handoff appends; the previous shift’s notes stay visible; the team builds a longitudinal record.
• Template enforced. The five sections (incidents, alerts, mitigations, planned changes, questions) are headings, not optional.

Antipatterns

• No handoff at all. Context lost.
• Handoff in chat. Lost in scrollback.
• 30-minute handoff. Wastes time; signal degrades.

What to do this week

Three moves. (1) Apply this practice to your next on-call rotation. (2) Survey the team after one cycle. (3) Iterate based on feedback; the discipline is the cadence.