Free to read ยท Three levels
Argo CD: Reconciliation, Ownership, and Delivery Decisions
Trace desired configuration through generation, comparison, sync policy, resource ownership, health, and compatible recovery.
Start learningIn this course
Eighteen developed lessons use a synthetic Parcel service to explain Argo CD control loops, Application and project scope, reproducible rendering, sync and health evidence, automation, apply options, hooks, custom health, Helm and Kustomize inputs, access control, ApplicationSets, multiple sources, diff exceptions, recovery, observability, and governance. Each lesson includes concrete examples, contextual mistakes, worked exercises, and official versioned primary references.
This course is a read-only resource. No repository is fetched, no Helm or Kustomize renderer is run, no Application or cluster object is applied, and no controller, job, service, credential, administrative interface, cloud operation, or deployment is contacted or started. YAML and Lua fragments remain runnable:false. Complete local Python programs evaluate declared synthetic data only and are verified in isolated temporary storage; they do not establish Argo CD or Kubernetes execution.
What you will learn
- Separate generation, comparison, sync operation, resource health, and user verification.
- Read Application and AppProject source, destination, kind, and authority contracts.
- Review automation, pruning, self-healing, shared ownership, and hook consequences.
- Trace effective render inputs, source vectors, generated-child ownership, and ignored-field scope.
- Maintain compatibility-aware recovery and bounded reconciliation observations.
- Evaluate source authenticity, target permission, and versioned policy independently.
Your learning path
Three levels
Desired state and evidence foundations
Read control loops, Application identity, rendering, projects, status, and sync review.
- GitOps reconciliation and the Argo CD boundary
- Application identity and desired-state specification
- Source generation and reproducible manifest inputs
- AppProjects and destination boundaries
- Sync status and resource health are separate observations
- Reviewing a sync before authorizing adoption
Automation, lifecycle, and trust
Review policy flags, ownership, hooks, health, render inputs, and scoped access.
- Automated sync, pruning, and self-healing
- Sync options and shared resource ownership
- Sync phases, waves, and hook side effects
- Custom resource health as a versioned contract
- Helm values and Kustomize transformation inputs
- RBAC, repository trust, and secret exposure
Generated systems and controlled recovery
Evaluate ApplicationSets, source vectors, diff exceptions, recovery, scale, and governance.
- ApplicationSet generation and child ownership
- Multiple sources and composite release identity
- Diff normalization and bounded ignored fields
- Rollback, desired-state correction, and recovery
- Reconciliation observability and measured scale
- Governance, source verification, and versioned change
Who this is for
- Developers learning to review GitOps desired state and adoption evidence.
- Platform and delivery maintainers reviewing ownership, policy, generated Applications, and recovery boundaries.
Before you start
- Basic Git, Kubernetes resource identity, YAML, application configuration, and testing knowledge.
- Ability to read small Python and JSON examples; no cluster, Argo CD account, cloud credential, or service is required.