Happy Monday everyone. Hope you have an amazing week. You got this!!

Breaking Through the First Wall Small win? Nah. This was a legit milestone. After wrestling through Google’s documentation, OAuth flows, and GCP setup, I finally pulled live JSON data from my Nest thermostat using curl. Seeing that blob show up in my terminal… felt like cracking a safe. What I learned along the way: - Navigating GCP I had to set up a full project in Google Cloud, enable APIs, link my Google Home account, and wire the whole thing together. This alone gave me a solid crash course in GCP’s structure and permissions model. - OAuth (finally clicked) Most of my past auth work has been SAML. OAuth has similar concepts, but the flow is different enough that you can’t fake it. Walking through the token exchange and scopes gave me a much clearer understanding of why modern API integrations work the way they do. - Real device integration → real data Once I completed the OAuth flow and hit the endpoint, getting that thermostat status back as JSON felt huge. That’s the “OK, now the real work begins” moment. What’s next? Now that I can authenticate and pull data, I’m starting the fun part: - Build small Python helper functions to clean and parse the data - Wrap it with FastAPI for a proper backend service - Store readings in Postgres - Visualize everything in Grafana - Run the whole stack in my local K3s cluster - Eventually test automated deployments + scaling This hits DevOps, APIs, cloud, auth flows, Python, infrastructure, and observability, all of the things I actually enjoy working on. What’s the coolest real-world device or API you’ve integrated into your homelab?

Just recently we saw what Gemini 3 Pro can do for websites. I decided that i have to use this to my advantage to build the Minecraft Server Management Dashboard of my dreams. I wanted to have my own project that i manage in a Development Environment and later send it to a Test, Production Environments to understand the big picture ideas. I got these ideas from a book i recently finished called The Phoenix Project, i'll make another post for it later in the future. So far i'm pulling modpacks via Modrinth API and be able to show every modpack's descriptions so i can understand what are their contents. Even though i'm gonna work on the website itself simultaneously, this week goals are: - On every push to Github repo, Github Actions runs tests and build docker images for backend & frontend individually then runs the containers on my server pc - Give it a domain name via reverse proxy with Nginx - Expose the metrics from containers and put them in Prometheus so i can watch them on Grafana I tweak and setup the website with AI and i like how it looks so far but the automation, the DevOps part excites me the most 😂 Can't wait to see how everything works by itself without me doing anything 😎 If you have any recommendations for DevOps structure, i'd love to hear them out. This will be my long term project for my homelab but i think it will be worth doing since it's so much fun!

I’ve been planning out what I want to build in my homelab, and the possibilities are… endless. Exciting and mildly overwhelming at the same time. Here’s where I finally landed: I’m using a Lenovo ThinkPad L13 Yoga (i3-10110U, 8GB RAM, 256GB SSD) running Ubuntu as the base for a K3s cluster. Not exactly a data center powerhouse, but perfect for learning Kubernetes fundamentals without smoking my electric bill. The Project Goal I’ve got a Google Nest (Gen 3) thermostat, and I want to pull real-world energy data out of it to analyze: - How often the furnace or AC runs - Average indoor temps - Seasonal patterns - Whether this thing is actually saving me money or just looks fancy Like I said, the laptop will be the cluster running multiple pods, and I plan to visualize everything in Grafana. I’ll figure out the guts in between. I learn best by doing — and doing something that actually interests me. IMO, this beats wandering through a course without a purpose. The Plan - FastAPI microservice to fetch & normalize Nest data - Store data in Postgres - Deploy everything in K3s (Deployments, Services, Ingress) - Visualize with Grafana - (Eventually) automate deployments via GitHub Actions → K3s Thoughts?

Today I had a small realization that might be useful for anyone getting into Linux. When you’re trying to pick your first distro or you just want to see what other ones offer, you don’t actually need to install them directly on your machine. Instead, you can just set up a simple virtual machine on your PC, test a few distros, and see which one fits your style. Personally, I never thought about doing that. Jumping straight into Arch seemed the most fun way because I enjoy the idea of building everything from scratch. But if you’re just starting out and want to test distros, setting up a VM first can save you a lot of time and make the choice easier. Just wanted to share this little tip in case it helps someone out here so you can be like the guy in the gif hahah