Argo Workflows

Information

Introduction

Argo Workflows is an open source container-native workflow engine for orchestrating parallel jobs on Kubernetes. Argo Workflows is implemented as a Kubernetes CRD (Custom Resource Definition).

Similar Software

Installation

ARGO_VERSION=4.0.2
#ARGO_VERSION=3.5.15
kubectl create namespace argo
kubectl apply --server-side -n argo -f https://github.com/argoproj/argo-workflows/releases/download/v${ARGO_VERSION}/quick-start-minimal.yaml
#kubectl apply --server-side -n argo -f https://github.com/argoproj/argo-workflows/releases/download/v%ARGO_VERSION%/quick-start-minimal.yaml

#?
kubectl patch deployment argo-server --namespace argo --type=json -p="[{\"op\":\"replace\",\"path\":\"/spec/template/spec/containers/0/args\",\"value\":[\"server\",\"--auth-mode=server\"]}]"
kubectl -n argo port-forward service/argo-server 2746:2746
#kubectl -n argo port-forward svc/argo-server 2746:2746
#kubectl -n argo port-forward service/argo-server 8080:2746
# Trust in Firefox Self signed certs from https://127.0.0.1
firefox --new-tab https://localhost:2746

argo submit -n argo --watch https://raw.githubusercontent.com/argoproj/argo-workflows/main/examples/hello-world.yaml

argo list -n argo
argo get hello-world-2qvr7 -n argo
argo get  hello-world-zglf9 -n argo
argo logs @latest -n argo
argo logs hello-world-2qvr7 -n argo
argo logs hello-world-zglf9 -n argo
argo delete hello-world-2qvr7 hello-world-zglf9 -n argo

#Same with kubectl:
wget -c https://raw.githubusercontent.com/argoproj/argo-workflows/main/examples/hello-world.yaml -O hello-world.yaml
kubectl create -f hello-world.yaml
kubectl get pods -n argo
kubectl get wf -n argo
kubectl get wf hello-world-2qvr7 -n argo
kubectl get wf hello-world-zglf9 -n argo
kubectl get po --selector=workflows.argoproj.io/workflow=hello-world-2qvr7 -n argo
kubectl get po --selector=workflows.argoproj.io/workflow=hello-world-zglf9 -n argo
kubectl logs hello-world-2qvr7 -c main -n argo
kubectl logs hello-world-zglf9 -c main -n argo
kubectl delete wf hello-world-2qvr7 -n argo
kubectl delete wf hello-world-zglf9 -n argo
kubectl delete pods -n argo -l workflows.argoproj.io/completed=true

argo delete --completed --all-namespaces

# Re-setup
ARGO_VERSION=4.0.2
# TODO : notes about stop, delete, ..
kubectl create namespace argo
kubectl apply -n argo -f https://github.com/argoproj/argo-workflows/releases/download/v${ARGO_VERSION}/quick-start-minimal.yaml

CentOS, Rocky Linux

Fedora

FreeBSD

OpenIndiana

Configuration

Usage, tips and tricks

Code Review Workflow

A specialized workflow for automated code reviews of multiple repositories can be found in the template and review directories. Specifically, the version in src/site/resources/argo/review/ is pre-configured for Minikube with host folder sharing.

The workflow performs the following steps:

  1. Setup: Creates a unique working directory clones/ to isolate the workflow execution.
  2. Clone: Clones the target repository into the unique workflow directory.
  3. Diff Generation: Generates a diff.patch for the repository using from/to commit hashes.
  4. Tasklist: Executes smi-build-tasklist on the cloned repository to generate TASKLIST.md.
  5. Review: Performs a non-interactive code review by executing the TASKLIST.md using the selected agent (claude or cursor), saving results to CODE-REVIEW.md.

Minikube Setup (Windows)

To use the code review workflow on Minikube with your local host folder, first mount the folder:

minikube mount C:\pub\setmy.info\data\minikube:/var/opt/setmy.info/minikube

Then manually clone your repositories into C:\pub\setmy.info\data\minikube\clones\. You can use the provided argo-wf.cmd script to automate cloning and submission:

cd src\site\resources\argo\review
# Usage: argo-wf.cmd <repo_url> <repo_name> <repo_branch> <from_commit> <to_commit> <agent>
argo-wf.cmd git@bitbucket.org:example/example-app.git example-app master HEAD~1 HEAD claude

If you need to clone multiple “secondary” repositories, you can create a file named secondary-clones.cmd in the same directory as argo-wf.cmd. If this file exists, it will be automatically called by argo-wf.cmd to perform additional cloning operations. Example secondary-clones.cmd:

@echo off
call clone-repo.cmd git@bitbucket.org:example/example-app2.git example-app2 master
call clone-repo.cmd git@bitbucket.org:example/example-additional-repo.git example-additional master

Apply the supporting resources from src/site/resources/argo/review/ (ensure 03-argo-secrets-map.yaml is updated with your ai-secrets):

kubectl apply -f src/site/resources/argo/review/00-argo-namespace.yaml
kubectl apply -f src/site/resources/argo/review/01-argo-host-path-persistent-volume.yaml
kubectl apply -f src/site/resources/argo/review/02-argo-host-path-persistent-volume-claim.yaml
kubectl apply -f src/site/resources/argo/review/03-argo-secrets-map.yaml
kubectl apply -f src/site/resources/argo/review/04-argo-role.yaml
kubectl apply -f src/site/resources/argo/review/05-argo-rolebinding.yaml

To submit this workflow with specific parameters:

# Basic submit (uses default HEAD~1..HEAD and agent=claude)
argo submit -n review src/site/resources/argo/review/argo-wf.yaml \
  -p uuid=$(uuid)

# Submit for a specific repo, commit range, and agent
argo submit -n review src/site/resources/argo/review/argo-wf.yaml \
  -p repo-name=my-repository \
  -p from-commit=v1.0.0 \
  -p to-commit=v1.1.0 \
  -p agent=cursor \
  -p uuid=$(uuid)

The default working-dir is set to /var/opt/setmy.info/minikube.

templates[n].nodeSelector for selecting concrete node type (OS, platform) from K8s to run images

See also