I’ve been having some hiccups on my HaOS, and then I started building yet-another monitoring dashboard for it. Of course, it created more questions than it solved
At first, I thought it was rebooting, but it’s just… hanging for the time it would take to reboot. Noticed that when I added
sensor.system_monitor_last_boot
to the dashboard and it was “10 days ago”. No signs of high CPU or RAM usage, nor temperature. But I digress.
Three minutes after updating HaOS (as a means to do a useful reboot), its swap hit 100% of usage. It only has 300 MB, and I never ever touched swap configs (didn’t even think about swap until I enabled the System Monitor integration, which I also never heard of before).
At 10 min of usage, RAM stabilized at almost 80%. Before the reboot, it was up for almost 10 days and was hitting 89% of RAM.
My add-ons are not far from basic ones, and no one seems up for blames.
I have lots of integrations, but I would say most are harmless.
A bunch of Zigbee devices, but I heard that Z2M is the memory hog, not ZHA which is what I use. I may want to move to Z2M at some point in the future (for better device support), but the amount of config kinda scares me.
I will move to a new house in the next year, and may end up adding more devices to the network - specially ceiling presence sensors.
Should I plan a hardware upgrade? Maybe a RPi4/5 with 2 or 4GB?
Or is this kind of usage expected and should be fine for a couple more years?
This guy is a tad old, had 2 years of mild usage as a retro-videogame and now is more than 3 years on the very same HaOS install
The peak you see in the main graph was the reboot. The holes on the smaller graphs below are the weird hiccups.