HomeAssistant Becomes Unreachable Every 12-24 Hours

Installation
1

In the last week or so, my home assistant has started becoming unreachable via the web UI every 12 to 24 hours or so. I’ve read through a few threads where people mentioning this happening but it seems not really match up to my symptoms at all.

My hardware is a Raspberry Pi 5, with SSD, running over POE, with 4G of RAM.

When the web UI becomes unreachable, I can still ping the machine successfully, but I cannot SSH to it (I have SSH running as an app on home assistant and can connect when everything is normal).

I also can still connect to http://homeassistant.local:4357/ and I see that supervisor, support and health are all showing as Healthy.

Rebooting the machine always seems to solve it, but then it goes offline again within a day.

I can tell when it becomes unresponsive because I have the System Monitor metrics enabled and at a certain point they go null and when I restart they start recording again so I can identify each outage timeframe.

Using the logs at http://homeassistant.local:8123/config/logs and I see no notable errors at the time that the machine becomes unresponsive.

My home network (which is ubiquiti unifi), shows the machine as still making network requests during the time that it’s inaccessible via the web UI as well, so it’s likely not totally dead (especially considering the observer URL works)

Does anyone have times on what else I should check to try to troubleshoot this? Is there additional logging I should consider turning on and if so, how?

Thanks so much for your help!

2

What time of day does this happen?
As you are powering with POE, did you try a different network port?
… Network cable?
Can you power the SSD separately to see if it is a power problem?

3

Have you recently added devices ? Based on my experience, your box may have reached its capacity, in which case you need to add memory.
I moved from a Yellow Box to a Mini PC and the issue was resolved.

4

It seems to happen at various times, here are the times of the last outages:

2026-08-28T11:30:26.536Z
2026-08-29T16:59:34.303Z
2026-08-31T01:44:05.081Z
2026-08-31T14:53:12.143Z
2026-09-02T19:56:00.479Z
2026-09-03T03:28:10.840Z
2026-09-03T06:25:36.474Z
2026-09-03T19:43:50.200Z

I haven’t added any new devices that I recall.

It is powered with POE, I’ll try a different port and different cable.

This is a RPI 5 with 4g of ram and a 250GB SSD which is hardly saturated. I don’t seem to ever fully use the RAM or CPU so I don’t think it’s a resourcing issue.

5

Pi 5 with SSD
POE
Ubiquiti
Argon box or SSD interface?
Just some trigger words that might put you on some search paths for clues.

6

Thanks, yeah I’m starting to suspect it’s a similar issue to this thread: Stability problems NVMe on RPI 5 - #53 by IOT7712

Although it’s weird that it only started happening in the last week or so, and I had been running for 6 months reliably before that.

7

H have the same issue using a RPI 4 and an external power supply. So I’d expect in your case it’s not PoE.

Last time when I discovered the issue I decided to troubleshoot later and after 30 to 60 minutes the connection recovered.

I am not sure if it is a software bug?
Until end of August beginning September I never experienced that issue.

I suspect that the HA server will reboot after it hung up.
Checking the Homeassistant Logs, I just see tons of fetching errors of my Unifi integration. I remved it, because I don’t use it. Let’s see if this is causing the issue.