Use this as an operational planning aid. Tailor checks to your application and approved change policy. A checklist is not a substitute for a tested disaster-recovery plan or contractual SLA.
Before action
- Identify service, owner, business impact, affected dependencies and incident priority.
- Confirm incident ticket, escalation path and current maintenance/change constraints.
- Gather logs, metrics and recent change evidence in a read-only manner.
- Confirm authentication, least-privilege access, backup status and rollback method.
- Identify and record the person authorized to approve any potentially disruptive step.
During controlled remediation
- Execute only an approved deterministic runbook or verified operator action.
- Restrict action scope to agreed hosts and services.
- Record operator identity, tool/job ID, timestamps and observed results.
- Stop and escalate if results diverge from expected preconditions.
Verification
- Verify application health and dependency connectivity, not just process state.
- Check monitoring stability and confirm service-owner acceptance.
- Record remaining risks, follow-ups, root-cause status and lessons learned.
- Reconcile the ticket, alert and change record; notify agreed stakeholders.
Recovery test planning
A restoration test must specify the test environment, recovery point objective (RPO), recovery time objective (RTO), data ownership, backups, service dependencies and restoration sign-off. Never run a destructive test against production without an explicitly approved procedure.