Code Room
On-callHard
Question
Your cloud audit-log anomaly detector pages at 02:50: an IAM principal belonging to a low-privilege CI build role performed a sequence in 90 seconds — iam:CreateAccessKey for an admin user, iam:AttachUserPolicy attaching AdministratorAccess to a newly created user, then sts:AssumeRole into the prod admin role and ec2:DescribeInstances across all regions. The build role should only be able to push to one ECR repo. No human was on shift. You're the on-call security engineer. Triage and respond.
What a strong answer looks like
Stop the bleeding first (mitigate), then form hypotheses from real signals. Separate root cause from symptom, communicate status as you go, and close with what prevents a repeat.
Learn the concepts
Loading whiteboard…
Run or narrate your approach, then ask the coach.