Hassio stops responding every few days / weeks (Proxmox VM)

I’m running Hass.io as VM on Proxmox (currently 0.93.1, had the same problem on earlier versions) with multiple add-ons, including Node-RED. Once in a while hassio stops responding - web interface is down, automations are are not running, but it is responding to ping and Telegram bot in Node-RED is still responding. I’m still learning and not using it for any production tasks, hence don’t always notice straight away that it is unresponsive. It runs at least few days, sometimes weeks between failures. After restart (from Proxmox) it comes back up without any problems. home-assistant.log starts at the boot time, so I’m not able to see what was happening before it froze (or if it was doing anything while frozen). During startup only some warnings regarding usage of custom integrations.
Tried googling around, but it seems typically such issues are due to Pi’s power supply, SD card failure or dropping wifi connection - all not applicable in my case.

Any ideas how can I troubleshoot this problem?

Thanks in advance,
Michal

Looks like a memory leak. Do you use any Netatmo components?

Not using Netatmo. I have grafana, influxdb, zigbee2mqtt (pretty new, issue appeared before zigbee2mqtt) and custom sonoff component.

I’m having a similar issue running Hassio via VirtualBox on my Mac mini. Every night it stops responding. Often times, just removing into my Mac mini and just hitting the return key in the VM, is enough to wake up Hassio. (I can confirm I’ve checked my energy settings and it’s not the Mac that’s going to sleep)—specfically the VM. Did you figure out what the issue was for you @Michal_Szymczuk?

Mine wasn’t (isn’t?) stopping that often, which makes troubleshooting more difficult. I didn’t have any issues in last 20 days, but was rebooting it quite a bit, so uptime wasn’t that long. I’ve got rid of custom Sonoff component (replaced switch with Shelly) and observing… if it will hang again I’m planning to stop various addons and see what happens.
Does your stop at the same hour every night? Some backup or other job running at that time?

Same Problem, VM also Turned OFF… try to start a full backup and 5min later its freeze and vm is down…

then i tried backup without frigate, media and influxdb and it worked… so i think its on of that three…