The first time a CloudTrail query saved us from a breach, it was already past midnight. The alert was subtle—just a sequence of API calls buried deep in logs. The account was valid, the credentials untouched. But something was wrong. And that’s when Zero Trust kicked in.
Zero Trust for AWS CloudTrail isn’t a philosophy. It’s a practice. A way to interrogate logs without blind spots, verify every action, and respond with precision. The days of trusting an internal IP range or a clean login are over. Every session is suspect, every event is tested against context. This is where CloudTrail Query Runbooks become your best weapon.
A well-built query runbook is more than saved SQL. It’s a living set of instructions—optimized to surface risk faster than any manual review. You start with patterns that matter: unusual AssumeRole calls, sessions from unexpected geographies, privilege escalations wrapped in innocent-looking API calls. Then you refine. Each runbook runs not once, but on repeat, automated, catching the faintest echoes of intrusion.
The power comes from pairing Zero Trust rules with targeted CloudTrail queries. You tag known accounts, segment workloads, and bake conditions into the SQL itself. You run the “list-sts” queries that expose role hopping. You check for IAM changes without ticket history. You track S3 access even when requests pass an ACL check. Every result is reviewed against business intent, not assumed safe just because it passed IAM policy.