Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Snapshot-based cloud forensics uses the platform’s own disk snapshot APIs to capture a compromised instance in seconds, then analyses the copy in an isolated forensic account.

What Is Snapshot-Based Cloud Forensics?
Snapshot-based cloud forensics is the practice of preserving a compromised virtual machine by calling the cloud provider’s own snapshot API rather than attaching a write blocker to a physical disk. The operation takes seconds, needs no access to the guest operating system, and produces a copy that can be analysed in a separate account while the original is left untouched or destroyed as the business requires.
It is the one area where cloud forensics is genuinely easier than the on-premises equivalent. Everything else about cloud investigation is harder, but acquisition is not.
Sequence matters more than tooling, and getting it wrong destroys evidence irrecoverably.
create-snapshots captures all volumes of an instance together, which matters for multi-volume consistency. Azure and Google Cloud offer equivalent managed disk and persistent disk snapshots.The straightforward route is to create a volume from the snapshot, attach it to a hardened analysis instance in an isolated VPC with no internet gateway, and mount it read-only. This is fast and familiar, but it means a live operating system is touching evidence, so mount options must be set carefully.
The cleaner route on AWS uses the direct block read APIs, ListSnapshotBlocks and GetSnapshotBlock, to stream the snapshot into a raw image file without ever attaching it. The result is an ordinary disk image that can be hashed, stored, and opened in Autopsy, X-Ways, or the Sleuth Kit exactly like any other. For evidential work this is preferable, because the image is immutable and its hash is meaningful.
Every API call is recorded by the platform itself. CloudTrail, Azure Activity Log, and Google Cloud audit logs show who created the snapshot, from where, and exactly when, which produces a stronger custody record than a paper exhibit form. Keep those logs for the duration of the investigation and export them with the evidence.
The preparation that makes this work is organisational rather than technical: a dedicated forensic account or subscription with a pre-agreed cross-account role, a key management service key reserved for evidence, and an automated capture function that an analyst can trigger without raising a change request. AWS publishes a reference solution for exactly this pattern, and the equivalent in Azure and Google Cloud is straightforward to build.
Where the compromised workload is a SaaS tenant rather than an instance, there is nothing to snapshot and the approach changes entirely; see SaaS evidence collection. We set up and rehearse both paths as part of Investigation and Incident Response Services.
No, not before capturing memory. Stopping or terminating an instance discards its RAM, taking with it running processes, decrypted payloads, and network state that often constitute the only evidence of a fileless intrusion. Isolate the instance at the network and identity layer first, acquire memory through the management agent, then snapshot the volumes while it is still running.
It is accepted practice, with caveats that should be stated. The snapshot is crash-consistent rather than a clean point-in-time image, and it has no inherent hash value. The usual approach is to materialise the snapshot into a raw image, hash that image, and rely on the provider’s own audit log to evidence who created the snapshot and when.
Because the account under investigation may still be under the attacker’s control. Copying the snapshot into a dedicated forensic account, re-encrypted with a key that only the investigation team can use, means that continued compromise of production cannot delete or alter the evidence. It also keeps investigative access separate from the operational permissions in the affected environment.
Files deleted from blocks that were previously written are usually recoverable, exactly as on a physical disk, because the data remains until the blocks are reused. Blocks that were never written at all are not stored in a thin-provisioned snapshot, so carving across unallocated space returns less than it would from a traditional image of a fully written drive.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.