Diary

Finished the DevOps school project (the AKS/Terraform/CI-CD one from August) and turned it into the actual fully formed file for the cert — writing up everything as a proper deliverable is somehow more work than building the thing in the first place.

Also did some spring cleaning on the esports side: r6-ops-board, gw2-ops-board and delta-ops-board got merged into one unified "Ops Hub", now living at ops.vaultinc.fr instead of three separate subdomains. Less to maintain, one site instead of three half-duplicated ones.

And the cluster grew a node — a Lenovo M70q Gen 6 (Core Ultra 5 225T) joined the Proxmox cluster. Still needs a job to do, but it's in.

Work is taking a lot of time and mental space, and the approaching exams too!

I may have fallen into Bannerlord for a while, but...

I (re)made this website! And the discord-1min.ai proxy & bot, and worked on the phone clusters. (There's a version of this exact website running on a OnePlus Nord N100 (like a backup) now!)

Also made the R6 Ops Board and the devops-feed-curator. All on docker, easier to maintain and update!

21/08

• Updated the r6-ops-board — now lives on my domain
• Updated the devops-feed-curator — now looks for deals (mini PCs/NUCs) and RAM sticks under a set price on eBay (via the eBay Dev Program)
• discord-1min — now works properly, has 9 different LLM models available and handles Discord thread conversations + internet search
• Started working on "delta-ops-board" to also follow Delta Force's competitive scene (still WIP — fork of r6-ops-board)

28/08

• Built "read-later" — paste a link, it fetches and summarizes it via 1min.ai. Lives at later.vaultinc.fr now, plain web form, no Discord bot needed for this one
• Also deep in the school DevOps project all week: Dockerized the ToDoList app, Terraform provisioning the AKS cluster, GitHub Actions CI/CD (looking at you, Azure OIDC), plus Prometheus/Grafana monitoring and HTTPS via cert-manager

I've worked on the homelab!

Mostly remodeling each LXC and services into... Yeah, services. Actual linux services, with restarts and errors, logs! And scripting the reboots & updates, because maintaining game servers when updates are weekly is getting a bit too much for my free time. So yeah, that. Simple stuff but effective, launching "reboot.sh" and watching it work is a vibe.

I was busy so there's really nothing to say for June. Oh, yeah, i made the Olivetin & Windrose servers and spent hours upon hours making it work. So, not much!

The general pattern behind pretty much every service on here: install Proxmox on the nodes, then every time I need something, spin up a new LXC and try to make it work. Generally it doesn't, first try. So I spend some time figuring out why (looking at you, Nginx and OliveTin), take some notes, poke at configs until something clicks, and then it works and I remember — mostly — how to fix it again next time it breaks.

Repeat that enough times and you end up with about 20 LXCs running various services, held together by notes, muscle memory, and stubbornness. Next up is automating a lot of this so I stop having to manually put my hands inside every stupid service that decides to stop working.