Reasons why I stopped using Azure and decided to use my home lab to build a portfolio and gain practical knowledge.
Well one reason mostly, is cost. Big risk of incurring 500 to 1000+ euros cost per month.
For reference, using the Azure pricing calculator, Just one Kubernetes cluster with 4 virtual machines, running 15 days (lets say i don’t have time every day to tinker with it) can cost 200+ euros! Not counting additional resources like subnet, load balancer, gateway, storage, possible ingress, egress data, monitoring, and other gadgets I’m forgetting right now.
For that sum, I can easily buy cheap hardware with 32Gb RAM, decent CPU, SSD storage, and run more than one cluster, non stop with any additional services I would want to experiment with.
Goal
My goal is to learn how things connect to form a production infrastructure in the real world. CI-CD pipelines, load balancing, security, monitoring health checks, database backups, and have a good practical understanding of the cloud environment around a containerized web app. Basically DevOps and Kubernetes.
I went through the AZ-400 Admin course. Found it difficult to follow, because of the very limited practical sandbox training. A lot of theory favoring power-shell commands but I’m used to Linux. The sandboxes started very slow or not at all. On top of that after 2 months in, somehow i got banned from opening the sandbox exercises and noticed a lot of people were receiving the same error.
Just reading and clicking through the lessons slides is too much theory and won’t persist in memory. I attempted a az-400 simulated exam after that and no surprise, did not went well.
Cloud test run
For more practice I opened an Azure account with free tier for one month, created some virtual machines, did a bit of networking, experimented with local Azure CLI and a few small Bicep templates. It was fun and that free month went by faster than you can type ip addr to view your network interfaces.
I was very careful to delete all resources to not go over the 200$ limit for the free tier.
In life you easily get distracted. Its easy to forget to switch off and delete the resources every time you have a spare one hour in the evening after work to learn.
After free tier expired, I was too stressed about accidental costs, considering the plan is to build a complete prod setup with ci-cd pipeline, monitoring, health check, load balancing and probably, Kubernetes later on.
My home lab
So as an alternative to cloud for now, the best thing is to use my existing home lab to see how far I can push it without needing additional hardware. I have OPNsense as a next gen router and a Mac mini M4 16gb RAM that should be enough to explore the simplest setup of clusters.
A home lab can be as complex as I can build it, no need to open it to the internet. I sleep well knowing i don’t have to remember to clean all resources, wait for them to be deleted and check again if i missed something still running outside of a resource group.
Creating virtual machines in UTM seems to work very well. Tested Vagrant with UTM plugin. After 3 commands I had a Ubuntu VM with working ssh connection, generated from a template file.
Of course Kubernetes at home is overkill when a simple container app will run just fine with software like Podman, Orbstack or Docker. My reasons for doing this are:
Proof of experience for career development
Trying out Kubernetes self healing and automated rollbacks on Home Assistant and other services running at home
The power of creating multiple VM’s and containers with one command
Next steps
Plenty to do next, but most important step is to generate 4 virtual machines that will provide base for a Kubernetes cluster, test how the Mini M4 16gb can handle multiple running virtual machines. So far it proved great for running Home Assistant in a VM.
Then will attempt to automate installing Kubeadm control-plane and nodes, try to run home assistant container version and follow how it behaves after connecting to a few devices.
That’s it for now, hope this will be part of a long series of Kubernetes journey.
Get in touch if you know devops communities. Would also love to connect and share experiences with other people running a home lab with Kubernetes.