Happy new year homelabbers! I’m cheating a bit and rolling some of the work I got done in December 2025 into this update. I want to say my homelab is pretty boring now it just works which is awesome, only minor tinkering is required these days so I’ve decided to focus on observability for a bit. Knowing when and what things have gone wrong was sorely missing so I’ve finally gotten round to sorting a lot of that out which I’ll detail in a section below.
Chicken/Egg Docker Woes#
Ever since I migrated away from Ubuntu to Debian I’ve been unable to restart the VM and have all the containers come up again. Even after going through every compose file and making sure that restart: unless-stopped was set. It’s been driving me nuts that this has been happening. Well I finally figure out what was wrong. The hint was that the only container to consistently come back up was fluent-bit. When I migrated to Debian I decided to improve my logging so set up fluent-bit and OpenSearch. I also set the docker logging driver to fluent-bit. Can you spot the issue yet? When docker comes up all the other containers fail over refusing to connect to the fluent-bit socket, pretty obvious now, if docker isn’t running the fluent-bit container isn’t running so how can docker connect to the fluent-bit container? I could keep this sentence recursive. Eventually the fluent-bit container comes up but the rest of them containers have already stopped, restarting them all brings them up fine. I removed the fluent-bit container and ran it as a systemd unit instead. Docker restarts are perfectly fine now.
Uptime Kuma and Gotify#
A while ago I did have Uptime Kuma as part of my homelab but removed it since everything was on one physical server so if there was a huge problem UK would never pick it up since it was probably down as well. I’m now hosting it on my Media Server which while still a single point of failure it’s physically separate from my main homelab. As well as this I setup Gotify to send me notifications when services go down, these go straight to my phone so I can get near realtime downtime notifications. I’ve had to tune the thresholds in UK for some of my applications since some are less reliable than others (the ones not on my infra!).
Wireguard on Android#
For about 24 hours I lost all network connectivity on my phone (Pixel 7 Pro). It turns out that wireguard had crashed and completely nuked connectivity. After rebooting my phone it worked again. No idea why this happened but it was pretty annoying. I usually have my home wireguard VPN connected 24/7 even when on my home wifi but because my phone is silent all the time I didn’t notice, I guess I didn’t need to use my phone at all that day.
That’s it for this update, there’s still work in my todo list that I need to get around to but that can wait for a bit.