No UI access but home is working

I have a weird problem my home assistant is basically running (lights work, automations work, Node Red flows all work) but I have no UI access at all :8123 is dead, :8437 is dead, I can’t access SHH (connects but the SSH side doesn’t respond before the client times out), although ping is happy to respond.

However, I can plug a monitor & keyboard in to access the CLI and running core check, supervisor check all seem to run fine.

Occasionally on a reboot the UI appears but crashes out within 2-3 minutes but not every reboot (more often the ones where rebooting is done with extreme prejudice).

Using the CLI the logs speed past so I can’t really use them, however going via login to the /mnt/data/supervisort/homeassistant/home-assistant.log and running cat with more/less or the tail -f command I can see are a large number of errors that keep flowing related to MQTT init.py and number to float conversions that start relatively soon after boot but nothing that is screaming or pointing to why web service & SSH has died.

I’ve tried to restore the most recent backup, I’ve tried downgrading core from the 2023.11.1 to 2023.10.5 via the CLI and that all appears to work fine but makes no difference to the symptoms

I’ve tried host reboot many times, core rebuild, & supervisor repair

But I am not seeing any change.

What should I be looking to try next to get this thing back up and running?

P.s. the hardware is Rpi 4 running from SDD using Home Assistant at the latest version

P.p.s I have a new SDD arriving today so I can try a fresh clean install and restore. However, I would like to try and get to the original working again as the backups have always excluded the history and influx databases as the sizes of them were causing problems with google drive backup and if I can I would like to keep that data.

The fresh install did exactly the same the second I restored a couple of days old backup.

However after a bucket load of head scratching I fixed it - it was the ZeroTier plugin that I had running which didn’t like a change I did the day before this started. The instance I figured out how to uninstall that plug in from the CLI everything suddenly sprang to life like nothing had happened.

Sodding networks