Skip to main content

Documentation Index

Fetch the complete documentation index at: https://notes.kodekloud.com/llms.txt

Use this file to discover all available pages before exploring further.

This lesson shows how to create and run a simple Argo Workflow, either from the Argo Workflows UI or with the Argo CLI. You’ll learn what the workflow YAML declares, how to submit it from the UI, how to submit via the CLI (including common flags), and how to inspect and retrieve logs for the run.

​
Overview

  • Argo Workflows are Kubernetes custom resources. You submit a Workflow manifest and the Argo controller creates pods to execute the defined steps.
  • Workflows are easily submitted from:
    • The Argo Web UI — great for exploring templates and visual debugging.
    • The Argo CLI — useful for automation and CI/CD pipelines.

​
Workflow YAML

Below is a minimal example Workflow. Key fields:
  • apiVersion / kind: identifies the resource as an Argo Workflow.
  • metadata.generateName: generates a unique name for each run, preventing collisions when rerunning workflows.
  • spec.entrypoint: the name of the template that should run first.
  • templates: one or more templates. Each template can be a container template and takes the same options a Kubernetes Pod container supports (image, command, args, env, volumeMounts, etc.).
apiVersion: argoproj.io/v1alpha1
kind: Workflow
metadata:
  generateName: cowsay-
spec:
  entrypoint: cowsay
  templates:
  - name: cowsay
    container:
      image: rancher/cowsay
      command: ["cowsay"]
      args: ["Losers visualize the penalties of failure, Winners visualize the rewards of success. -Dr.Rob Gilbert"]
generateName is helpful when running the same workflow repeatedly: it appends a unique suffix so each workflow run gets a distinct name and avoids resource-name collisions.

​
Submit from the UI

  1. Open the Argo Workflows web UI and click “Submit New Workflow”.
  2. Paste the YAML above (or select a saved template) into the submission form.
  3. Create the workflow. The UI will show a summary (name, pod name, host node, phase, and action buttons like Resubmit, Suspend, Stop, Terminate, Delete).
  4. Click into the workflow to view templates, inputs/outputs, and logs.
A screenshot of the Argo Workflows web UI showing a workflow named "cowsay-7pzj4." The right-hand summary panel lists details like ID, pod name, host node, phase "Pending" with message "PodInitializing," while control buttons (Resubmit, Suspend, Stop, Terminate, etc.) appear across the top.

​
Submit from the CLI

The Argo CLI supports many flags for connecting and authenticating to the Argo Server. Below is a concise table of common flags and their purpose:
FlagPurposeExample
--serverAddress and port of the Argo Server--server argo.example.com:2746
-n, --namespaceNamespace for the CLI request-n argo
--kubeconfigPath to kubeconfig (only needed for out-of-cluster)--kubeconfig ~/.kube/config
--tokenBearer token for Argo Server authentication--token $ARGO_TOKEN
--username, --passwordBasic auth credentials--username user --password pass
-e, --secureUse TLS (defaults from env var)--secure
--insecure-skip-tls-verifySkip TLS verification (insecure)--insecure-skip-tls-verify
--watchStream workflow progress until completion--watch
-v, --verboseEnable verbose logging (debug)-v
A more complete (version-dependent) list of flags is available from argo --help. Check the Argo components and other pods in your cluster (example using kubectl): 
kubectl -n argo get pods
NAME                                  READY   STATUS      RESTARTS   AGE
argo-server-68bc7d7567-468lc          1/1     Running     0          3m59s
httpbin-58d7595979-xndwm              1/1     Running     0          3m59s
minio-5cb4ff75c9-stmmw                1/1     Running     0          3m58s
workflow-controller-5bb95ffc45-rp2fz 1/1     Running     0          3m59s
Submit the workflow from a remote URL and stream progress with --watch: 
argo submit https://gist.githubusercontent.com/sidd-harth/54c1c6e16682ae281650a7d67bf0bf01/raw/7cae878eaad64b84a486841bb9ee484cea83ca45/workflow-1.yml -n argo --watch
When you submit a workflow using generateName, Argo assigns a unique name (for example, cowsay-nzs5t). The CLI will stream the workflow progress and show the final status.

​
Inspecting the workflow

You can get a concise workflow summary from the CLI (or inspect via the UI). Example argo get output (abridged): 
Name:                 cowsay-nzs5t
Namespace:            argo
ServiceAccount:       default
Status:               Succeeded
Conditions:
  PodRunning          False
  Completed           True
Created:              Fri Oct 24 06:28:18 +0000 (34 seconds ago)
Started:              Fri Oct 24 06:28:18 +0000 (34 seconds ago)
Finished:             Fri Oct 24 06:28:52 +0000 (now)
Duration:             34 seconds
Progress:             1/1

STEP          TEMPLATE   PODNAME         DURATION  MESSAGE
✓ cowsay-nzs5t  cowsay     cowsay-nzs5t   23s
Check pods again to see workflow-created pods and their status: 
kubectl -n argo get pods
NAME                                   READY   STATUS      RESTARTS   AGE
argo-server-68bc7d7567-468lc           1/1     Running     0          32m
cowsay-7pzj4                           0/2     Completed   0          2m23s
cowsay-nzs5t                           0/2     Completed   0          44s
httpbin-58d7595979-xndwm               1/1     Running     0          32m
minio-5cb4ff75c9-stmmw                 1/1     Running     0          32m
workflow-controller-5bb95ffc45-rp2fz  1/1     Running     0          32m

​
Viewing logs

You can retrieve logs using either kubectl or the Argo CLI: kubectl logs (pod): 
kubectl -n argo logs cowsay-nzs5t
Sample output (kubectl logs): 
time="2025-10-24T06:28:37.839Z" level=info msg="capturing logs" argo=true

/ Losers visualize the penalties of \
| failure, Winners visualize the rewards |
\ of success. -Dr.Rob Gilbert            /
-----------------------------------------
         \   ^__^
          \  (oo)\_______
             (__)\       )\/\
                 ||----w |
                 ||     ||

time="2025-10-24T06:28:38.841Z" level=info msg="sub-process exited" argo=true error="<nil>"
argo logs (workflow): 
argo -n argo logs cowsay-nzs5t
Sample output (argo logs prefixes output with pod name): 
cowsay-nzs5t: time="2025-10-24T06:28:37.839Z" level=info msg="capturing logs" argo=true
cowsay-nzs5t:
cowsay-nzs5t:  / Losers visualize the penalties of \
cowsay-nzs5t: | failure, Winners visualize the rewards |
cowsay-nzs5t: \ of success. -Dr.Rob Gilbert            /
cowsay-nzs5t: -----------------------------------------
cowsay-nzs5t:
cowsay-nzs5t:   \   ^__^
cowsay-nzs5t:    \  (oo)\_______
cowsay-nzs5t:       (__)\       )\/\
cowsay-nzs5t:           ||----w |
cowsay-nzs5t:           ||     ||
cowsay-nzs5t: time="2025-10-24T06:28:38.841Z" level=info msg="sub-process exited" argo=true error="<nil>"
The same logs are available in the Argo UI when you open the workflow and click “Logs”. At this point you should see two workflows in the UI (one created from the UI and one created via the CLI), both producing the same cowsay output when you inspect their logs.
A browser screenshot of the Argo Workflows web UI showing a workflow named "cowsay-nzs5t" with status, start/finish times, duration, conditions and labels. The left sidebar displays namespace and filter options and the top bar has "Submit New Workflow" and "Completed Workflows" buttons.

Watch Video

Practice Lab