6 comments

[ 3.5 ms ] story [ 24.2 ms ] thread
I think there is a mistake:

> Pod: The smallest kubernetes unit: it can run multiple pods inside of it

Should be: "it can run multiple containers inside of it".

I remember trying to set-up a kubernetes cluster on a weekend for similar use case and I also didn't have much background in k8s, although bit more than OP. I had used it before, created deployments, set-up secrets etc. But when I tried to set it up at my new work place, I failed.

In our use case, we didn't have to account of unknown number of users. We had an airflow instance which needed to run jobs in parallel and I didn't have much idea about celery either. Finally I created a docker swarm cluster which was much easier to set-up and fulfilled our requirements.

IMO there is zero point to running a non-managed K8s cluster anymore unless you are at enterprise/fang scale and have a full team of CKAs to babysit it. Pay AWS or GCP the small fee for their managed services and worry about the kube API and the things that make your company money.
> To add ArgoCD to your cluster you just simply type

    kubectl create namespace argocd
    kubectl apply -n argocd -f https://raw.githubusercontent.com/argoproj/argo-cd/stable/manifests/install.yaml
If that makes Kubernetes simple, then everything is simple, since someone somewhere wrote a script that you can curl | bash.
I feel like this is an unfair comparison. A more correct one would be compiling binaries from source vs using a Makefile.