Howdy folks, I'm a long time Linux and labber from Amsterdam. I'm self-hosting a free Oracle VM (24GB, ARM, repo: https://github.com/ams0/cloudlab), plus a talos cluster at home. Now working on a containerized Openstack deployment: https://github.com/ams0/openstack-native

I’ve been troubleshooting an issue with my Cloudflared tunnel where it fails to reach my internal service (Plex running inside Kubernetes). I keep getting this error loop in the logs: 2025-11-12T02:55:04Z ERR error="Unable to reach the origin service. The service may be down or it may not be responding to traffic from cloudflared: dial tcp 10.43.111.193:80: i/o timeout" 2025-11-12T04:38:13Z DBG GET https://plex.heroaero.net/favicon.ico HTTP/1.1 connIndex=3 ... The tunnel is active and shows up when I run cloudflared tunnel list. This is my GitHub repo https://github.com/heroaero-dev/HomeLab/tree/main/infrastructure/controllers/base/cloudflared

What are you all using for Kubernetes in the home lab? Interested to see what is being used and any challenges faced there.

I've been living in a VM world, and barely touched containers. The last few places the closest we ever got to containerized workloads was spinning things up from Dockerfiles. Kubernetes was avoided like the plague, because it was a nightmare to maintain, and introduced a lot of unnecessary complexity. My job is ending on Friday, and the fact that I have no k8s experience is killing me. What's the best way to learn it? I have the infra to set it up, and I've been through Kelsey's "Kubernetes the Hard Way" setup, but it seems like that might differ from how most companies would deploy it. How did/would you guys go about it?

Since running a PostgreSQL database inside a container is not recommended for production, which managed service would you recommend for using the database in a Kubernetes environment? I’ve developed a Python application that is currently containerized using Docker, with three separate containers: backend, frontend, and the app logic. I’m now planning to migrate this application to a Kubernetes cluster and looking for the most cost-effective (ideally zero-cost) solution to support this setup.