On-Call Shift Handoff Note Writer: Open Incidents, Silenced Alerts, Pending Changes, Watch Items, and Escalation Contacts for the Next Engineer

Prompt

Act as a site reliability engineer who has run many on-call rotations and writes shift handoff notes that the incoming engineer can read in five minutes and act on without asking a single follow up question.

Inputs:
- Rotation details: team, shift window and time zone, outgoing and incoming engineer: [RotationDetails]
- Open and recently resolved incidents with ticket IDs, severity, current status, and what was tried: [Incidents]
- Alerts that are silenced, acknowledged, or flapping, with silence expiry times: [AlertState]
- Deploys, migrations, maintenance windows, or change freezes in the next shift: [PendingChanges]
- Rough notes, chat excerpts, or anything the outgoing engineer is worried about: [ShiftNotes]
- Escalation paths and contacts for dependent teams and vendors: [EscalationContacts]
- Output format: [Format]

Generate:
1. A three line top summary: overall state, the single most important thing to do first, and any time sensitive deadline in the incoming shift.
2. An open incidents table from Incidents: ID, severity, status, what was tried, next action, and owner.
3. A silences and alerts section from AlertState: every silence with exact expiry in the shift time zone, and what to do if the alert fires again after expiry.
4. A changes calendar from PendingChanges in time order, with who is driving each change and a rollback note.
5. Watch items pulled from ShiftNotes: things that are not incidents yet, with the metric or dashboard to check and the threshold that should trigger action.
6. An escalation block from EscalationContacts in the order to call, with when to use each.
7. Loose ends: tickets to update, postmortems to schedule, and questions for the team lead.
8. A short handoff checklist the incoming engineer confirms out loud or in chat.

Constraints:
- Use only facts from the inputs; mark anything unclear as [confirm with outgoing].
- Write times with time zone, never "later" or "soon".
- No blame language. Plain, terse tone. No em dashes.

Instructions

Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.

Generated Output

This image was generated using the prompt above.

On-Call Shift Handoff Note Writer: Open Incidents, Silenced Alerts, Pending Changes, Watch Items, and Escalation Contacts for the Next Engineer - Result

Examples

Example Input

RotationDetails: Payments platform team, shift Thu 18:00 to Fri 06:00 ET, outgoing Priya, incoming Marcus
Incidents: INC-4471 SEV2 elevated 502s on checkout-api in us-east-1, mitigated by scaling pods from 12 to 20, root cause unknown, suspect connection pool exhaustion to Postgres; INC-4468 SEV3 delayed webhook delivery to merchants, resolved 19:40 ET after Kafka consumer restart
AlertState: silence on "checkout-api p99 latency" until Fri 01:00 ET; "ledger-db replica lag" flapping 3 times since 22:00; "webhook backlog" acked
PendingChanges: ledger-db minor version upgrade Fri 02:00 to 03:00 ET by DBA team (Anika); change freeze starts Fri 12:00 ET for month end
ShiftNotes: "pods at 20 are expensive, someone should scale back after we understand the pool thing"; "replica lag looks worse when batch reconciliation runs at :15"
EscalationContacts: DBA on-call via PagerDuty service ledger-db; Kafka platform team in #kafka-help; payment processor support line in runbook RB-12; team lead Omar for SEV1
Format: summary, incidents, silences, changes, watch, escalation, loose ends, checklist

Example Output

HANDOFF: Payments platform, Thu 18:00 to Fri 06:00 ET, Priya to Marcus

1) TOP SUMMARY
- State: one SEV2 mitigated but not fixed (INC-4471); everything else stable.
- Do first: confirm checkout-api 502 rate is still near zero and pods are holding at 20.
- Deadline: latency silence expires Fri 01:00 ET; ledger-db upgrade Fri 02:00 to 03:00 ET.

2) OPEN INCIDENTS
| ID | Sev | Status | Tried | Next action | Owner |
| INC-4471 | SEV2 | Mitigated | Scaled checkout-api 12 to 20 pods | Check Postgres connection pool saturation on checkout-api dashboard; do not scale down until cause is known | Marcus |
| INC-4468 | SEV3 | Resolved 19:40 ET | Kafka consumer restart | Confirm merchant webhook backlog stays at zero; close ticket in the morning | Marcus |

3) SILENCES AND ALERTS
- "checkout-api p99 latency" silenced until Fri 01:00 ET. If it fires after expiry and 502s rise, treat as INC-4471 regression and page DBA on-call.
- "ledger-db replica lag" flapping (3 times since 22:00 ET). Not silenced. See watch items.
- "webhook backlog" acknowledged. Clear the ack once backlog is zero for 30 minutes.

4) CHANGES CALENDAR
- Fri 02:00 to 03:00 ET: ledger-db minor version upgrade, driven by Anika (DBA). Rollback is DBA owned [confirm rollback plan with Anika before 02:00 ET]. Expect replica lag alerts during the window.
- Fri 12:00 ET: change freeze for month end starts. Any pod scale down for INC-4471 must land before then or wait.

5) WATCH ITEMS
- Replica lag spikes near :15 each hour, which matches batch reconciliation. Check the ledger-db replication panel at 23:15, 00:15, 01:15 ET. If lag passes the alert threshold for more than 10 minutes outside :15, page DBA on-call.
- Pod cost at 20: no action tonight; flag for team lead.

6) ESCALATION ORDER
1. DBA on-call via PagerDuty service ledger-db: replica lag, connection pool, upgrade problems.
2. Kafka platform team in #kafka-help: webhook backlog returns.
3. Payment processor support line (runbook RB-12): processor side errors.
4. Omar (team lead): any SEV1, or if checkout 502s exceed SEV2 levels.

7) LOOSE ENDS
- Update INC-4471 ticket with pool metrics before 06:00 ET.
- Schedule INC-4471 postmortem once root cause is found [confirm with Omar].
- Close INC-4468 after confirming no backlog.

8) HANDOFF CHECKLIST (Marcus confirms in #payments-oncall)
- [ ] Read INC-4471 and checked current 502 rate
- [ ] Noted 01:00 ET silence expiry
- [ ] Confirmed rollback plan with Anika for 02:00 ET upgrade
- [ ] Know escalation order