CAPA Study Plan¶

Five weeks at 5-7 hours per week. Weight the first two weeks toward Argo Workflows, which is the largest and least familiar domain.

Week 1: Argo Workflows fundamentals¶

• Install Argo Workflows on kind and run the hello-world example
• The Workflow custom resource and the controller's role
• Template types: container, script, resource, suspend
• Orchestration templates: steps versus DAG, and when each fits
• Entrypoint, templateRef, WorkflowTemplate, ClusterWorkflowTemplate
• Input and output parameters, and passing values between steps
• Lab: build a DAG with four tasks and a parameter flowing through
• Review Notes: notes/01-argo-workflows.md

Week 2: Argo Workflows advanced¶

• Artifacts: repositories, input and output artifacts, artifact garbage collection
• Volumes and volumeClaimTemplates for sharing data
• Conditionals with when, loops with withItems, withParam, withSequence
• Retry strategy, timeouts, activeDeadlineSeconds
• Exit handlers and lifecycle hooks
• CronWorkflow, and the concurrency policy options
• Pod garbage collection, TTL strategy, and the workflow archive
• Synchronization with mutexes and semaphores
• Service accounts and workflow security
• Lab: a CronWorkflow with an artifact, a retry strategy, an exit handler, and a TTL

Week 3: Argo CD¶

• Architecture: API server, repo server, application controller, Redis
• The Application resource: source, destination, project, syncPolicy
• Sync policy: manual versus automated, prune, self-heal
• Sync waves, hooks (PreSync, Sync, PostSync, SyncFail), and ordering
• Health status and custom health checks
• ignoreDifferences and fields owned by other controllers
• AppProject for multi-tenancy: allowed sources, destinations, kinds
• ApplicationSet generators: list, cluster, Git, matrix, merge, pull request
• Kustomize, Helm, and plugin support
• App of apps and declarative Argo CD management
• Lab: an Application with automated sync and self-heal, then an ApplicationSet with a list generator
• Review Notes: notes/02-argo-cd.md

Week 4: Argo Rollouts and Argo Events¶

• Rollout as a Deployment replacement, and how it manages ReplicaSets
• Canary strategy: steps, setWeight, pause, and traffic routing
• Blue-green: active and preview services, autoPromotionEnabled, scaleDownDelay
• Traffic providers: Istio, NGINX, ALB, SMI, Gateway API
• AnalysisTemplate, AnalysisRun, and metric providers
• Automatic promotion and rollback behavior
• Experiments
• Argo Events: EventSource, Sensor, EventBus
• Event sources, filters, triggers, and payload parameterization
• Lab: a canary Rollout with a Prometheus analysis that fails, and a webhook Sensor triggering a Workflow
• Review Notes: notes/03-argo-rollouts.md and notes/04-argo-events.md

Week 5: Integration and review¶

• Trace an end-to-end path: event, workflow, commit, sync, rollout
• Know which project owns which responsibility, and where boundaries lie
• Work every scenario in scenarios.md
• Two timed practice exams
• Re-read the Workflows documentation on templates and artifacts

Readiness check¶

• Explain the difference between steps and DAG templates and when each is right
• Describe how a parameter and an artifact each pass between workflow steps
• Explain what prune and self-heal do in an Argo CD sync policy
• Describe when to use an ApplicationSet rather than several Applications
• Explain sync waves and give a case that needs them
• Describe a canary rollout with analysis and what happens when analysis fails
• Explain the roles of EventSource, Sensor, and EventBus