Web UI unavailable after a while – Only fixed by ha core restart (since 2025.7.1)

Installation Home Assistant OS
1

Hello everyone,

Since I upgraded to Home Assistant 2025.7.1, I’ve been facing a strange issue:

  • The web interface becomes unreachable after some time (a few hours).
  • A manual ha core restart brings it back — but only temporarily.
  • All background services appear to continue running normally.
  • I can ping the device, access via SSH, and run core commands.

This is happening on a system running Home Assistant OS 15.2 (machine: qemux86-64) with Supervisor 2025.06.2.

Here are the relevant diagnostics I ran:

:white_check_mark: Core Status (while the UI is inaccessible):

$ ha core status
blk_read: 0
blk_write: 0
cpu_percent: 4.94
memory_limit: 8329433088
memory_percent: 12.64
memory_usage: 1052700672
network_rx: 0
network_tx: 0

$ ha info
arch: amd64
channel: stable
docker: 28.0.4
homeassistant: 2025.7.1
hostname: homeassistant
operating_system: Home Assistant OS 15.2
machine: qemux86-64
state: running
supervisor: 2025.06.2
supported: true

$ ha network info
interfaces:
	•	interface: enp0s18
connected: true
ipv4:
address:
	•	192.168.3.11/24
gateway: 192.168.3.1
method: auto
nameservers:
	•	192.168.3.1
ready: true
host_internet: true
supervisor_internet: true

I can ping everywherez from HAOS :

$ ping -c 4 192.168.3.1
64 bytes from 192.168.3.1: time ~0.4 ms

$ ping -c 4 8.8.8.8
64 bytes from 8.8.8.8: time ~15 ms

$ ping -c 4 google.com
64 bytes from google.com: time ~15 ms

:no_entry_sign: Problem summary:

  • After some uptime, the web interface (http://192.168.3.11:8123) becomes unavailable.

  • No visible CPU or RAM spikes.

  • network_rx and network_tx show 0 in ha core status during this time.

  • Everything comes back with → ha core restart

  • UI then works again… until it freezes hours later.

Possible lead:
I suspect the core process becomes unresponsive on port 8123 while the rest of the system stays alive. The issue appeared only after updating to 2025.7.1 — never happened before.
SSH remain usable, everything seem to work in the background (automations, Z2M, …)

:speech_balloon: Has anyone else experienced this issue since 2025.7.1?
Any idea if this is a known bug in the release?

Thanks in advance!

2

I‘m having the same issue

3

I might be seeing the same issue. Every so often I catch the app or the HA tab in my browser showing the “reconnecting” message similar to what it shows when HA is being rebooted. Is this what you see as well? I’ve had this issue for a while now.

4

That ‘reconnecting’ message comes from this setting :

afbeelding

1 Like
5

I was incorrect, the message says “Wrapping up…” which is the same one you get after a reboot.

6

Same here. With 2025.7.2 i have the same problems. After a while Web UI is unavaiable. Any Updates?

  • Core2025.7.2
  • Supervisor2025.07.1
  • Operating System16.0
  • Frontend20250702.2
7

Did you ever get this sorted? I have been having the exact same issue since 2025.9.1. I even rolled back a version and the issue has stayed.

8

Have you checked whether HA still has network access? In my case, webUI goes down but the console is up, and HA simply has no network connection caused by a NIC driver crash (the e1000 hang issue). Since my instance runs in Proxmox and after changing HA to use a different NIC than the one needed for host management, I am able to bring down (already is but I run the command anyway) and back up the NIC used by HA and access to the webUI is restored. Could your issue be the same? Before finding it was a NIC issue, I too could only fix it with a reboot.