Skip to content

Progressive Delivery Home Lab

Level:Intermediate
Time:2-3 hours
Cost:low
Works locally:Yes
Cloud creds:No
Cleanup:Yes
Reviewed:2026-05-30
Validation:make validate

Local Kind or Minikube lab for canary releases and rollback.

Outcome Kubernetes resource output plus app/GitOps screenshot
Tools used Kubernetes, Argo Rollouts, Progressive Delivery, Canary Deployment
Best fit Intermediate - 2-3 hours
Student workstation Repository files Lab steps Local Kubernetes Validation proof
  • Install or review: Kubernetes, Argo Rollouts, Progressive Delivery, Canary Deployment.
  • No cloud provider credentials are required by the project metadata.
  • This project can be practiced locally before you publish portfolio evidence.
  • Open the safety guide before running commands that create infrastructure.
Cost and credential stance

Cost risk is low. Cloud target: no cloud provider. Cloud credentials needed: No. Always use your own account, never commit secrets, and confirm cleanup before creating paid infrastructure.

Use this flow before you run commands:

  1. Read the cost and credential warning above.
  2. Review the validation, troubleshooting, cleanup, and portfolio proof sections below.
  3. Follow the original project guide preserved near the bottom of this page.
  4. Return to the validation and cleanup checks before you capture portfolio evidence.

Run the project validation command before and after meaningful changes:

Terminal window
make validate
  • Run make validate first so local tooling issues are visible early.
  • If a command fails, check tool versions, working directory, and required environment variables.
  • For Kubernetes failures, check the current context, namespace, pod events, and service endpoints.

Cleanup is available or expected for this lab. Use the cleanup or destroy steps in the guide below, then confirm that local clusters, containers, cloud resources, buckets, state files, and CI secrets are no longer active.

  • Validation command output: make validate
  • Screenshot or terminal proof: Kubernetes resource output plus app/GitOps screenshot
  • Notes explaining what changed, what failed, and how you fixed it
  • Cleanup evidence, especially for cloud or Kubernetes resources

The original README content is preserved here for lab-specific commands and context. Headings are intentionally demoted so the page outline stays focused on the standard lab flow.

External image reference Level
External image reference Cost
External image reference Cloud
External image reference Delivery

Local Kubernetes lab for learning canary releases and automated rollback with Argo Rollouts.

  • Why progressive delivery is safer than all-at-once deploys
  • How canary steps work
  • How service selectors route traffic during rollout
  • How to pause, promote, and abort a rollout
flowchart LR
Student["Student changes image tag"] --> Manifest["rollouts/rollout.yaml"]
Manifest --> Controller["Argo Rollouts controller"]
Controller --> Canary["Canary ReplicaSet"]
Controller --> Stable["Stable ReplicaSet"]
Service["Kubernetes Service"] --> Canary
Service --> Stable
  • Docker
  • Kind for the one-command workflow, or Minikube if you prefer manual setup
  • kubectl
  • Argo Rollouts kubectl plugin, optional but recommended
Terminal window
make validate
make up
make logs
make down

make up creates a local Kind cluster named rollout-lab, installs Argo Rollouts, and applies the sample rollout.

Terminal window
kind create cluster --name rollout-lab
kubectl create namespace argo-rollouts
kubectl apply -n argo-rollouts -f https://github.com/argoproj/argo-rollouts/releases/latest/download/install.yaml
kubectl wait --for=condition=available --timeout=180s deployment/argo-rollouts -n argo-rollouts
kubectl apply -f rollouts/
kubectl argo rollouts get rollout demo-rollout -n progressive-delivery --watch

Change the image tag in rollouts/rollout.yaml, apply it again, and watch the canary steps.

Terminal window
make validate

This parses the namespace, service, and rollout YAML locally, so it works before a cluster or Argo Rollouts CRD exists.

  • no matches for kind "Rollout": install Argo Rollouts before applying the manifest to a cluster; make validate is only a local YAML check.
  • Rollout never progresses: check pod readiness with make logs, then inspect the image tag in rollouts/rollout.yaml.
  • Argo Rollouts plugin missing: use kubectl get rollout -n progressive-delivery for a basic status view.
  • Kind cluster already exists: run make down, or use CLUSTER=my-rollout-lab make up.
Terminal window
make down

For Minikube users:

Terminal window
minikube delete
  • Add an analysis template.
  • Introduce a bad image tag and abort the rollout.
  • Change canary weights.
  • Compare this with a normal Kubernetes Deployment.
Use the guide first.

The full learning flow stays on this page. Open GitHub only when a step asks you to inspect code, fork the lab, or download source assets.