GCP Compute Snapshot Created by Rare User
Description
Identifies the first time, within a lookback window, a principal creates a snapshot of a Compute Engine persistent disk ("*.compute.disks.createSnapshot", "*.compute.regionDisks.createSnapshot", or "*.compute.snapshots.insert"). Adversaries create disk snapshots to stage data for exfiltration, copying or sharing the snapshot to an attacker-controlled project without reading the source VM directly. This is a New Terms rule keyed on the project and acting principal, excluding known cloud-provider egress networks.
Query · kuery
data_stream.dataset: "gcp.audit"
and event.action: (
*.compute.disks.createSnapshot or *.compute.regionDisks.createSnapshot or *.compute.snapshots.insert
)
and not event.outcome: "failure"
and cloud.project.id: *
and user.email: *
and not source.as.organization.name: (
"Google LLC" or "Amazon.com, Inc." or "Microsoft Corporation" or "GOOGLE-CLOUD-PLATFORM"
or "MICROSOFT-CORP-MSN-AS-BLOCK" or "AMAZON-AES" or "AMAZON-02"
)
Investigation fields
Pivot points the source recommends for triage.
@timestampuser.emailsource.ipsource.as.organization.namegcp.audit.resource_nameevent.actionevent.outcomecloud.project.id
Known false positives
- Backup, disaster-recovery, and image-automation service accounts create snapshots routinely. Baseline known automation principals by user.email and add exceptions; interactive or first-seen snapshot creation warrants review.
Analyst notes
Disclaimer: This investigation guide was created using generative AI technology and has been reviewed to improve its accuracy and relevance. While every effort has been made to ensure its quality, we recommend validating the content and adapting it to suit your specific environment and operational needs.
Investigating GCP Compute Snapshot Created by Rare User
GCP persistent-disk snapshots are point-in-time copies of a zonal or regional disk created via *.compute.disks.createSnapshot, *.compute.regionDisks.createSnapshot, or *.compute.snapshots.insert. Adversaries snapshot a target disk and then copy or share it to an attacker-controlled project, enabling data theft without touching the source instance.
Possible investigation steps
- Identify the acting principal (
user.email) and whether snapshot creation (event.action) is expected for it. - Review the target resource (
gcp.audit.resource_name) and request parameters to identify the source disk and whether it holds sensitive data. - Review
source.ipandsource.as.organization.name; confirm the operation aligns with normal egress. - Correlate with follow-on snapshot copy/share or image creation in the same project or across projects.
False positive analysis
- Backup/DR and image-bake automation; baseline those service-account principals and add exceptions by
user.email.
Response and remediation
- If unexpected, review any snapshot copy/share, revoke cross-project access, rotate the implicated principal's credentials, and tighten least privilege on
compute.disks.createSnapshotandcompute.snapshots.create.