A practical cloud security review
Review cloud changes, permissions, alerts and backup exceptions with named owners. Check logging coverage and follow through on unresolved findings.

A cloud review is useful when it produces a decision: investigate an unexpected change, restore a failed protection or confirm that a known exception is still needed. Opening several dashboards and seeing green indicators isn’t enough.
Choose the systems the business depends on and name the person who reviews each one. Agree how urgent alerts reach someone outside the review period. The scheduled check sits on top of day-to-day monitoring and response.
Start with responsibilities and coverage
List the accounts, subscriptions and applications included in the review. Record which organization manages access, configuration, patches, backups and incident response for each service.
Microsoft’s shared-responsibility guidance distinguishes virtual machines from managed platforms and ready-made applications. Hosting a VM in Azure still leaves operating-system and application management with the customer. Data and identity protection remain customer responsibilities across service models.
Check that the reviewer can see every relevant environment. A clean report from one subscription says nothing about an account missing from that report.
Match changes to a change record
Review new administrator assignments, access-policy changes, storage exposure and changes to monitoring or backup settings. Match them to a change record and the responsible owner.
In Azure, the activity log records management operations. Reads and actions inside a resource need different logging; resource logs require configuration. Verify the needed logs are being collected and retained before relying on them to investigate access.
For an unexplained change, record the resource, time, initiating identity and immediate concern. Pass it to the person responsible for investigating. Save the logs and other evidence before changing settings, but follow the incident plan if something needs containing right away.
Check who can reach business information
Look at effective access, including public endpoints, shared links, guest accounts, service identities and inherited permissions. Start with changes and sensitive systems, then schedule a broader access review.
Azure Blob guidance explains how account-level settings can prevent anonymous container access. There are other ways in, so review authenticated permissions, shared access signatures and separately configured website features too. Microsoft’s Blob security guidance identifies the static website endpoint as a separate consideration.
Some resources intentionally serve public content. Record their purpose and owner, and keep private records separate. Don’t close a production endpoint without checking what depends on it and planning the change with its owner.
Investigate exceptions with context
Review high-priority alerts, unexpected sign-ins, failed backup jobs and overdue updates. Compare unusual resource use with deployments, scheduled jobs and expected business activity. High CPU use alone doesn’t identify malware or an attack.
For each finding, record an owner, priority, next action and follow-up time. If it looks like an active compromise, use the incident process immediately instead of leaving it for the next review.
Check backup completion against the agreed schedule and recovery needs. Investigate a failed job before repeatedly restarting it. Keep the results of restore tests separately, since a successful job won’t tell you whether the application can be recovered.
Close the loop
At the next review, confirm that open findings were resolved and test the protection that had failed. Give every exception an end date and revisit it when the service changes.
Tune alerts that produce repeated noise, and test that an important alert reaches the intended person. Record gaps in logging, support hours or ownership as work to complete, with someone accountable for it.
Use the separate cloud-cost review for spending decisions. Talk to us if you need help defining cloud review coverage and the handoffs that follow a finding.


