ArgoCD Starter Guide: Full Tutorial for ArgoCD in Kubernetes

Want to learn GitOps? ArgoCD is GitOps for Kubernetes. In this practical tutorial video I teach you everything you need to know get started with ArgoCD on your Kubernetes cluster. Watch or follow along with your own Kubernetes cluster and let's implement ArgoCD!

📘 Chapters:
00:00 What you will learn in this video
0:30 Helm Theory - Everything you need to know
1:33 Example Pipeline for ArgoCD
2:45 ArgoCD Disaster Recovery and Multi-Cluster Support
3:36 What we will do in the Lab
4:15 Installing ArgoCD, Portforward, Argo Password, Login
6:39 Deploying a Helm Application with ArgoCD
15:12 ArgoCD to Rollback applications
17:48 Deploying a Kustomize application with ArgoCD
19:44 Kustomize vs Helm Configmap Demonstration
25:13 Using ArgoCD CLI to create, sync, delete and troubleshoot

#Kubernetes #ArgoCD #GitOps #Tutorial #GettingStarted #Declarative #Configuration #ContiuousDelivery #Deployment

[DevOpsJourney]

Let me know what technology you are trying to learn next

[MK-yx]

Thank you, I think this is the only tutorial which explains rollback process clearly.🎉

[kazwalker]

Around 25:00, when using Helm, you should use a checksum annotation of the config map in the deployment. Google "helm tricks automatically roll deployments", it'll be the first result from the Helm docs.

Great video, cheers!

[ayencoscolfield]

this was a delight to watch absolutely brilliant presentation on Argocd, thanks for the effort

[matk]

Excellent Tutorial on ArgoCD !! - Absolute pleasure to listen to you. Gratitude !!

[anantharamakrishnan]

😍The Video. Clear concise explanations. I knew Kustomize before and I always wondered why config generator. Now I get the reason. Thanks for your explanations.

[sonnix]

great simple and clear tutorial, thank you sir.

[JasonBerk]

I'd love to see a video that covers bootstrapping and using Argo CD from the ground up without using the UI

[nguyenhoangphuckan]

This video was so helpful. Thank you for making it!

[Maksat....]

Great tutorial! Thank you so much, Brad!

[sudo-bandit]

Can you do another real time DevOps CI/CD lab? Really enjoyed your last one and found your mentorship insightful!

[Shrivanurag]

great tutorial to begin with. clear and concise.

[jalandharbehera]

Its too very good informative videos. thank you so much sir ❤

[RockyScenes]

Regarding the destination cluster URL - if there were more than one, what would they look like? How do we define them so we know which one is which? I am guessing that one shows as default because it was not configured with a different name?

[TAsSos]

Thanks a lot for this nice and easy guide regarding ArgoCD! The only that wasn't clear to me is the best way to update our app. Actually I have exactly this issue in my environment. I have to change every time the Helm Chart version if I want to actually deploy a new change on configMaps. So I'm looking for a way to fix this. If we have no a new release but some configuration changes how to setup ArgoCD to update the application directly if it detects just a simple change in the configMap.

I will go now to watch also your videos regarding the Helm chart and I'm curios to see the differences between ArgoCD and Helm cli, since argocd looks like it is able to do what exactly helm client can do and much more.

[webentwicklungmitrobinspan]

15:15 what happens to the users that are interacting with one of the replica that are beeing terminated?
do they notice something?
i have asked ai about it but i am not convinced that it will be that graceful. i am thinking of the applications i am developing with a few 100 parallel requests taking ~20sec to complete each..
i wounder if its safe during usage hours.

claude says:
When changing replicas from 5 to 2 in Argo CD with auto-sync enabled, users interacting with your application might notice some effects, though Kubernetes tries to minimize disruption:
What happens technically:

You commit a change to your Git repository, updating the replicas from 5 to 2
Argo CD detects this change (typically within 3 minutes by default)
The auto-sync feature automatically applies this change to your cluster
Kubernetes selects 3 pods to terminate (reducing from 5 to 2)

What users might experience:
If you have proper load balancing:

Minimal or no disruption
Requests in progress on the terminating pods will complete (Kubernetes sends SIGTERM and waits for graceful shutdown)
New requests are routed only to the remaining pods

Potential issues users might notice:

Increased latency: The remaining 2 pods may experience higher load until autoscaling kicks in (if configured)
Session loss: If you're not using persistent sessions or session affinity, users whose connections were on the terminated pods might need to log in again
Temporary capacity reduction: During the transition, you might have reduced capacity to handle traffic spikes

Best practices to minimize disruption:

Use PodDisruptionBudgets to ensure a minimum number of pods remain available
Implement proper health checks and readiness probes
Set appropriate termination grace periods
Use session persistence mechanisms if your application maintains user state

For critical production workloads, you might want to consider using Argo CD's manual sync approval for replica count changes to have more control over when the scaling occurs (perhaps during low-traffic periods).

[tejit]

Hi, at 10:37, you are saying one argocd installation can manage multiple kubernates cluster, but i don't understand how? because argocd is going to be installed in a specific cluster, so is there not one-to-one relationship between argocd and K8S cluster, if not how you can manage multiple K8S cluster?

[lbaibiAbderrahman]

we should apply this cmds on terminal or just cmd

[TaeglicheNarzisst]

Getting a "Unable to create application: error while validating and normalizing app: error validating the repo: error verifying source type: multiple application sources defined: Helm, Directory" upon first hit of "Create" around Minute 10.00..

[MuhammadKamranAzeem]

Thanks for the video, but apparently the most important piece is missing. When the actual application code changes, how do you build a new image (with a new tag other than "latest"? and how do you update the helm and kustomize configurations with that new image name:tag, and how do you trigger anrgo CD that application code has been changed, new image has been created and it (argo cd) should kick off and deploy the newest version? In your video you only showed what happens when the configuration part changes. So it is incomplete IMHO.