Agree what recovery must achieve.
Identify the business activities supported by each application and the effect of losing recent changes. Recovery point objective concerns the point in time to which data must be recovered. Recovery time objective concerns the tolerated recovery duration. The two objectives influence different parts of the design.
Prioritise databases, files, configuration, identity services and external dependencies. Document which data can be recreated and which records cannot. A cloud disaster-recovery plan also needs a viable destination, capacity, licences and the people authorised to restore normal service.
Choose protection that fits the failure scenarios.
Set backup frequency, retention, protected components and storage separation explicitly. Consider accidental deletion, corruption, account compromise, ransomware and loss of the primary environment. A copy accessible through the same compromised account may offer limited protection against a malicious deletion.
Immutable or retention-locked storage can be considered where supported, with its exact retention settings, override permissions, recovery access and cost reviewed. A snapshot can assist a short-term recovery task, but its independence and protection depend on the platform. Confirm what remains available if the source service or account is lost.
Operate the backup process and its exceptions.
Backup management can include job checks, failure investigation, capacity review and retention changes within the agreed scope. Define who notices a failed job and what action follows. Distinguish an incomplete copy, an application-consistency issue and a warning that needs investigation.
Protect credentials and encryption keys separately enough to support the intended recovery scenario. Document who can restore, delete or change retention. Include logs and test copies in data-location and retention discussions. A protected dataset still needs controls over the people and processes that can access it.
Test restoration and maintain the runbook.
Practice restoring representative data and validate the application, permissions and dependent services. Record the usable recovery point, elapsed stages and manual interventions. High availability may reduce interruption from some failures, while backup and recovery address different loss scenarios.
An agreed BLCS Global scope can produce a protection schedule, responsibility matrix, recovery runbook and test record. Share the platform, data volumes, current backup method and required recovery objectives when enquiring. Availability of specific storage controls and any recovery commitment are confirmed for the selected service.
