Incident simulations
Exit

INC-3016 · P2

No deployment goes through any more: the CI gets “You must be logged in to the server (Unauthorized)”

Casablanca · AWS account eu-west-3 · Amane Plus Assurances · 15:30

Amane Plus Assurances (a fictional car and home insurer) deploys its claims portal on EKS (cluster eks-amane-prod). Until last week, Jenkins deployed with the IAM role jenkins-deploy.

The DevOps team has just migrated to GitHub Actions with OIDC federation: a new role gha-deploy-eks was created, with a policy that allows eks:DescribeCluster. Jenkins was shut down this morning.

An urgent fix (wrong deductible calculation on car claims) must ship today. The GitHub Actions job fails at the “Déployer” (deploy) step.

You are on your workstation with the AWS CLI v2 and kubectl with the cluster administration role (eks-admin), region eu-west-3. The gh CLI is connected to the amaneplus/portail-sinistres repository.

This simulation is reserved for students enrolled in a course.

Already enrolled? Log in.