Establish the change timeline
For every reboot, routing change, DNS update, boot switch or intervention, record the request time, completion time and expected result. Correlate that interval with monitoring and application logs. The operation history proves what the platform attempted; the service test proves whether the intended result was achieved.
Investigate a failed operation
Capture the operation name, timestamp, target resource and complete error text. Then check the server's power state, network reachability and application health independently. A failed maintenance task does not necessarily mean the server is down, while a completed task does not prove the application recovered.
Distinguish scheduled and completed work
Treat planned or pending work as a future change, not as current configuration. Confirm that a scheduled operation is still required before its window begins and cancel it when the original reason no longer applies. After execution, verify the resulting boot, network or service state directly.
Prepare a maintenance record
Document the objective, affected services, customer impact, start and end criteria, rollback path and responsible person. Preserve relevant operation identifiers and times for the incident record, but never copy console credentials or other secrets into it.
