SOC 2 Evidence from AWS CloudTrail
Updated
Why this matters for SOC 2
If your production infrastructure runs on AWS, CloudTrail is the primary evidence source for almost every infrastructure-related SOC 2 control — who made what change, who assumed what role, and whether access followed least-privilege principles. Auditors will ask for it by name.
What evidence CloudTrail provides
- A complete, append-only record of every IAM change (CC6.3 — role-based access and least privilege)
- Evidence of system monitoring for anomalous API activity, supporting CC7.2
- A record of console and API access patterns that can demonstrate access reviews are grounded in actual usage, not just policy documents
- Cross-account access and AssumeRole activity, which is frequently the specific evidence requested for multi-account AWS organizations
How LogTriage maps this to SOC 2 controls
IAM-modifying CloudTrail events are mapped to MITRE ATT&CK privilege escalation techniques, which the compliance mapper then ties to CC6.3 (Role-Based Access and Least Privilege). Detected exfiltration-pattern activity maps to CC7.2 (System Monitoring) automatically, with the specific evidence note an auditor expects attached to the report.
Evidence checklist
- Confirm CloudTrail is enabled in every account and region in scope, including organization-level trails
- Verify log file integrity validation is enabled — auditors specifically check for this
- Retain evidence of IAM access reviews and tie them to actual CloudTrail-observed usage
- Document the retention period for CloudTrail logs and confirm it meets your SOC 2 evidence window requirement
- Maintain a record of any privilege escalation alerts and their resolution
Frequently Asked Questions
- Which CloudTrail event types are most important for SOC 2 evidence?
- Events related to IAM changes (user creation, policy attachment, access key creation), sign-in events from signin.amazonaws.com, and any errorCode: AccessDenied on sensitive APIs. These directly satisfy CC6.1 (access controls implemented) and CC6.3 (access restriction for sensitive activities).
- Does SOC 2 require CloudTrail to be enabled in all AWS regions?
- Yes, if your infrastructure spans multiple regions. SOC 2 auditors expect coverage to match your actual AWS footprint. An organization-level trail or multi-region trail is the typical implementation. Single-region CloudTrail in a multi-region deployment is a finding.
- How long do we need to retain CloudTrail logs for SOC 2?
- At least the audit period (typically 12 months), plus enough prior history to establish a baseline for anomaly comparison. Common practice is 1 year in the primary S3 bucket and longer in cold storage for forensics purposes.
- Can a single LogTriage report serve as SOC 2 evidence, or do auditors want raw logs?
- Typically both. The LogTriage report demonstrates that logs are reviewed, anomalies are detected, and findings are documented — satisfying the monitoring component. Raw CloudTrail exports or S3 log references satisfy the log retention and availability component. Neither alone is complete.
Related Resources
Reference
- AWS CloudTrail Log Format: field reference → every field explained, with real sample records
- AssumeRole — A request to AWS STS for temporary credentials for an IAM role; the record names the caller, the role and the temporary access key it returned, which ties the role session to every call made with it.
- AWS CloudTrail events: ConsoleLogin, AssumeRole, CreateAccessKey, StopLogging… →
- Paste one record and see what each field means → free, runs in your browser
See your compliance mapping generated automatically
Every LogTriage report includes a deterministic compliance mapping — SOC 2, PCI DSS, HIPAA, NIST CSF, and ISO 27001 — stamped on every report, AI-generated or rule-based.