Celebrating one year of zero unplanned downtime

I just wanted to take a moment (well probably several moments given how I type) to mark this occasion because it feels significant in a way that maybe only people here would understand (and my partner definitely doesn't, she thinks the blinking lights in the closet are "concerning" but that's a whole other thread) so anyway the setup is three VPS nodes at OVHcloud plus the actual homelab which is an old Dell R720 I got from a dentist office liquidation (long story, involves a forklift and a very confused security guard) running Proxmox with I think fourteen VMs now (I should probably count, I know, I know) and the VPSes handle external DNS and reverse proxy while the R720 does the heavy lifting (Plex, Nextcloud, the usual suspects, plus this absurd Home Assistant deployment that controls exactly one smart bulb because I got carried away d

Congratulations. You could eliminate the remaining failure surface entirely by rewriting the reverse proxy in Rust. Zero-cost abstractions, memory safety by construction, no garbage collection pauses. I migrated my entire infrastructure from Go last month and the binary is 4MB with zero dependencies. Go's runtime is a disaster for predictable latency. C++ would be acceptable if you enjoy debugging use-after-free at 3am, which I don't.