A good answer is a good start.

Monitoring detects conditions and sends notifications. Response, automation and on-call ownership need separate definition.

Start with critical user journeys, service access and failure indicators such as full disks. Tune alerts to avoid unnecessary noise.

Share architecture, key metrics, budget range, goals and constraints. Sensitive access credentials are not needed for an initial review.

Depending on scope, outputs can include an architecture proposal, options comparison and implementation steps. Agree deliverables at the start.

Results depend on overwritten data, media condition and failure type. Recoverability cannot be guaranteed before assessment.

Stop unnecessary writes to the affected environment. Record errors, recent actions and the last working time for assessment.

Configuration review and penetration testing have different scopes. Define methods, targets and authorization at proposal stage.

Reporting and applying fixes may be separate work. Agree response, retesting and deliverables at the start.

Operating-system maintenance and application development are different. State supported services and application boundaries in the offer.

Use scoped accounts and appropriate secure access methods. Do not put passwords in public requests or email messages.

Snapshots often share dependencies with their source platform. Assess an independent backup with separate failure domains and recovery policy.

Define acceptable data-loss time, then choose frequency around change rate, retention cost and restore needs.