Suddenly my solution ran for almost48 hours without crashing but the day after it stopped twice within 24h.
I have tried almost everything.
New SD, New Raspberry PI, checked memory, temperature, CPU load, reduced number of integrations.
No help
Seems like many have the same problem and it existed for a loooong time.
Rebooting every 6 hours is not an acceptable option.
Does anyone have an idea, how we can escalate this?
I’m having the same problem with RPi 4, SSD setup. I have tried disabling various plugins, addons, restarting nightly and changing the recorder config. I still get random freezes every few days and sometimes within 24 hours of unplugging/replugging. As @JanRoman mentioned, I am also seeing “the green LED at the ethernet port blinks very rapidly and very regularly” when this happens.
At my system it was able to kill my router completely. All lan ports and wan port went down. The pi4 also became totally irrisponsive, blinking the network port like crazy.
In the end I did try for a few days to remove the cap so a little more cool air came at the pi board. Perhaps you can try that too. But did not wait for another crash.
After 2 weeks struggle I decided to move to a nuc. Having a very stable system.
I hope you can find the answer.
And after upgrading to SSD and adding more air do even decrease temperature. Same issue.
Dows anyone have a good idea on how we can increase the priority on this topic?
Why are only some having this?
Did you find the solution? I have exactly the same problem as you.
Ok, as suggested by a few others on this thread, I did a completely fresh install of HA and did not restore from a Snapshot. So far, it has been stable! It is a bit of a pain setting everything up again by hand but I did extract one of the snapshots locally and search through it for settings config, etc. to make the setup easier. Although it’s a pain, having HA lock up every 24-48h is definitely more of a pain.
Another variable at play is that when I set things up again, I used the Z-Wave JS Integration rather than the old Z-Wave one. I have an unconfirmed suspicion that old integration was causing issues but I do not know for sure.
Im having the same problems but dont use z-wave at all… have tried stripping most of the add-ons but cant figure out any source of the problem. 
Have you tried a completely fresh install without restoring a snapshot?
I have tried restoring a snapshot, and even a fresh install. I don’t think it is a problem of HA itself, but I believe it is some kind of low level Hass OS problem as there simply is no pattern to it whatsoever. Is there anyone on Debian for instance experiencing something like this?
I’m having this issue too…
Normally i wipe the DB and its ok again but now no luck wiping the db
The system goes unreachable after x time
What should i do?
Yesterday evening, all was ok, this morning had to reboot
Fine for an hour, now back to unreachable
Pls
Help me 
Hi.
I’m facing the same issue since about 2 weeks, RPI 3B+ and SSD on USB.
Everything started after installing Esphome, but I’m not sure that things are related.
Same issue, running RPI3 HassIO.
What should have been a fun project has become a nightmare.
I suggest moving away from SD cards for a start. Yes the RPi is ‘suggested’ in the docs, but that is purely to people a way of getting into HA for very little financial investment.
Perhaps you are following extremely outdated YouTube videos from when Home Assistant was still called this? It hasn’t been HassIO for something like 18 months or more. Anything you watch or read that still calls it this is surely not accurate enough to still use and will likely cause you trouble.
Hi
I followed the latest and greatest installation procedure for RPI, i just assumed it was Hassio.
I even tried to install HA on Debian 10, but the behaviour was exectaly the same. RPI 3 became frozen in matter of minutes or few hours.
However, long story short, i went ahead and got myself a brand new Raspberry Pi 4 - 8 GB with a decent SD card, and installed everything from scratch again, and it has been running fine for almost 3 days now, smooth as butter, (knock the wood) 
Im currently running these apps:
Nah, HassIO hasn’t existed (as that name) for a long time. Long story short, its’ the same product but a new name, now simply called Home Assistant.
Unfortunately we still keep getting people use the old name in forum posts.
I am experiencing the same problem with a Raspberry Pi 4b 4GB with an application class 2 SD card. Home Assistant no longers serves the UI over port 8123, and I cannot ping the device on the local network. Connecting a HDMI cable and keyboard to the Pi reveals that the system is indeed online and I can process commands (like host reboot which fixes the problem). Is there some sort of log file I can sift through to see what might have happened during the time when it became unresponsive? Did the web server crash? Or was it something more sinister…
Couple of ways to see log data from the terminal:
dmesg => reavels kernel diagnostic info.cat /config/home-assistant.log => HA Logstop => CPU and Mem usage.Worth a shot
But all these will be wiped out on boot, so you should try to read these files while the system hangs.
Thanks for the advice. I did have the problem again today and only thought to check back here after the fact.
This time I had it slightly different, the UI would respond okay, but every integration had stopped working. Logs seemed to indicate some type of network error that I failed to make a note of. I thought I’d try to restart using the web interface using Configuration → Server Controls, but the instance never came back online. I rebooted from the terminal and it seems okay again now.
I am sure this will happen again in a few days, so I’ll try to run those commands on the box to see what it shows.
Just for the record.
After MANY months of problems with state of the art RPI4 setup and all sorts of reinstall etc. I finally decided to run in on an older Intel NUC - No problems ever since.
This is definitely HassOS related (at least using RPi4). As many other users reported, all problems stopped after migrating from HassOS to Debian. I even restored the backup, no problem whatsoever.