On-call operations

An on-call shift connects a service to a responder, with start and end times and an escalation level. The incident records its assigned person separately. That distinction matters when a service has an on-call plan but a specific case is assigned to another responder. The Monitoring Operator can consult the shift, assign the incident and record the first page event. The responder acknowledges and records the action taken. If the case changes hands, later response events show who handled the escalation.

Suppose the database service fails an availability check near a shift change. The first responder is on the current shift, while the next responder starts soon. The operator opens the incident and pages the current responder. That responder acknowledges and writes an investigation update. If work continues beyond the shift, the team updates assignment and records an escalation event for the person receiving it. The chronology should show the page, acknowledgment and next handoff at their real times. A shift boundary by itself does not resolve an incident.

• Coverage. Keep shift start, end, service and responder together for the team reviewing an alert.
• Escalation. Use the escalation level and a response event to record the actual handoff.
• Incident history. Preserve attributed updates so the incoming responder can see what has already been tried.

Agree which services require scheduled coverage and who owns shift maintenance. Set escalation levels in language the team uses and choose a procedure for overlapping or uncovered shifts. Decide how the monitoring operator confirms the receiving responder before changing assignment. The person listed in a shift may be unavailable; the actual response should be clear from the incident and event records, not inferred from the rota alone.

Set up adjacent shifts for one service and create a failed check near the handover. Ask the Monitoring Operator to identify the current responder, open the incident and record a page. Have that responder acknowledge, add an update and hand the unresolved issue to the next person. The incoming responder should be able to read the preceding action and add their own update. The Service Owner then checks whether there is one current assignee and whether the response events identify each actor.

Repeat with an escalation to a higher level. The shift record should explain who was expected to take the page, while the event records show what actually happened. If a page received no acknowledgment, the operator keeps the incident open and escalates through the agreed path. Recovery is recorded only after the service check and impact evidence have been reviewed.

• On-Call Schedule — See service coverage by start, end, responder and escalation level.
• My Incidents — Find work assigned to the current responder.
• Open Incidents — Review cases waiting for acknowledgment or further investigation.

Use incident response for assignment and acknowledgment. See outage assessment when the case affects service availability.