Kubernetes? Does anyone here use it?

I use ClusterAPI for bootstrapping clusters (using local Kind cluster as management cluster).
Cluster powered by ClusterAPI are then on GCP or local on premises clusters (we do have an internal implementation for a ClusterAPI infrastructure provider).

To deploy add-ons I use projectsveltos https://github.com/projectsveltos (FD, I am one of the maintainer of projectsveltos)

Calico as CNI and Istio as service mesh, then Prometheus.

Hi!
Not sure about ur use case but.. For personal projects I use two setups depending on my needs:

• a managed K8s on DigitalOcean (rather cheap options as I don't have any heavy services yet)
• self-managed single node K3s for sandbox projects/playing around
  From my experience getting into K8s might be a bit time consuming and troublesome and it might take some time to setup a working cluster. BUT - once you have one - it speeds the development a lot (at least for me) and things like CI/CD as well.

IMO after getting the basic building blocks right - it's not that difficult to manage a K8s project - lot of people like to demonize K8s as very difficult to maintain - it doesn't have to be like this if you keep things simple and don't have too much custom/special features (like I have in my daywork but that's a humongous enterprise project).

Regarding stack I think the following is nice to have and it's my personal choice for bootstrapping a cluster:

• some API gateway, like Gloo, Kong - currently I'm going to test KrakenD but I think they all have some pros and cons - once setup it makes things easier.
• it's nice to have a service mesh set up - Istio is my choice, I don't know any others
• use helm whenever I can - or at least do everything as code if possible to keep track on changes and reuse them easier.
• ArgoCD is nice have
• Certificate-manager for SSL certificates
• nb, I prefer to keep the DBs outside K8s ;)
• kind/minikube for local development
• kubeadm if you chose to self manage the cluster

I'd expect hosting costs to be higher, at least for small/side projects - it usually changes when the project grows (where K8s actually shines imo).

Managed Kubernetes cluster with OVH.

No extra cost. We pay for instances hosting out node pools. The cost of nodes is the same as plain vanilla public cloud instances.

We created a complete stack for Kubernetes that adds developer- and operations-centric tools, automation and self-service on top of Kubernetes in any infrastructure or cloud. 1 install, 1 spec, and 1 console to build, deploy, secure, expose and monitor containerized applications. Check it out on Github: otomi-core