version: 1
id: incident-capstone
title: Capstone — Restore the Checkout Platform
summary: Triage a multi-layer release failure and leave an auditable incident record.
difficulty: capstone
estimatedMinutes: 60
prerequisites: [linux-navigation, git-recovery, docker-debugging, docker-networking, kubernetes-deploy]
image: alpine:3.21
shell: /bin/sh
setup:
  - apk add --no-cache git
  - cd /workspace && git init -q && git config user.email learner@platformforge.local && git config user.name Learner
  - "printf 'services:\\n  checkout:\\n    image: checkout:latest\\n    environment:\\n      REDIS_URL: redis://localhost:6379\\n  redis:\\n    image: redis:7-alpine\\n' > /workspace/compose.yaml"
  - "printf 'apiVersion: apps/v1\\nkind: Deployment\\nmetadata:\\n  name: checkout\\nspec:\\n  replicas: 1\\n' > /workspace/deployment.yaml"
  - cd /workspace && git add . && git commit -qm 'broken release'
tasks:
  - id: contain
    title: Repair the release configuration
    description: Pin checkout to 1.4.2, target the redis service, and scale the Kubernetes deployment to three replicas.
    hints:
      - Fix the service hostname before changing anything else.
      - Replace moving image tags with the supplied release version.
    checks:
      - type: command
        name: Checkout image is pinned
        command: "grep -q 'image: checkout:1.4.2' /workspace/compose.yaml"
      - type: command
        name: Redis service discovery is correct
        command: grep -q 'redis://redis:6379' /workspace/compose.yaml
      - type: kubernetes
        name: Checkout has three replicas
        command: "grep -q 'replicas: 3' /workspace/deployment.yaml"
  - id: evidence
    title: Record and commit the incident outcome
    description: Create incident.md with sections Root cause, Mitigation, and Follow-up, then commit all changes.
    hints:
      - Use Markdown headings so the record is easy to scan.
      - Confirm git status is clean after committing.
    checks:
      - type: file
        name: Incident record identifies root cause
        path: /workspace/incident.md
        value: Root cause
      - type: command
        name: Incident record includes mitigation and follow-up
        command: grep -q 'Mitigation' /workspace/incident.md && grep -q 'Follow-up' /workspace/incident.md
      - type: command
        name: Recovery is committed
        command: cd /workspace && test "$(git rev-list --count HEAD)" -ge 2 && test -z "$(git status --porcelain)"
limits: {cpus: "0.75", memory: 192m, pids: 96, timeout: 3600, network: true}