I have this issue intermittently. Sometimes a few minutes, sometimes hours. None of the logs give a hint to what it is. No CPU/RAM spike, no network outages. SSH works. Ping to HA works. Disabled most of plugins.
Today it’s not reachable for more than 6 hours. However sensors send their data to the backend. Only UI does not seem to work properly.
Please give me some hint what I can try, I’m desperate. Did a ha core restart from SSH with negative result.
From outside or inside your home network? When it is not reachable, can you ping it? Have you tried from multiple devices?
If from the outside: share how you made your HA accessible.
Also share:
- what kind of setup you are running
- versions
- have you installed/updated some 3rd party integrations
- any difference if you start HA in safe mode
Yes, multiple devices, PC, Android, iPhone. From within using IP or from outside (I use duckdns with SSL). Nothing works. I can ping successfully. Also can connect via SSH internally and use ha console. Everything normal there. But UI is not reachable. This is now for half a year. Sometimes it’s just a few minutes a day, sometimes it’s hours. After some time it just works again as if nothing happened.
Raspberry PI with SSD running home assistant OS.
DuckDNS extension with HTTPs. Adguard extension.
Zigbee connector per USB.
All latest updates installed.
Have this now for half a year or more… getting frustrated as my setup is growing.
Could it be a certificate issue?
Intermittently? This should be then coincidentally with IP renewal? Does not seem so.
Also sometimes it’s not working for hours other times it’s not working for a few seconds to minutes but dozens of times over the day.
How could I debug this?
Yeah just throwing ideas out there. So this also happens if you access it directly by IP:8123 internally it’s probably not a certificate issue. I would start by testing with duckdns disabled, see if the issue still pops up.
You are certain you don’t have any IP conflict on your network I assume.
Thanks for providing ideas, very much appreciated. no conflicts in network, raspberry gets static IP.
Yes I try IP:8123 with same negative result.
Edit:
Need to correct myself. This time I’m not able to connect using https://IP:8123 BUT http://IP:8123 works. That’s however not each time the case. It seems the longer outages are coincidental with http working.
I would firstly try to find out if connection locally with HTTP://HA_IP_ADDRESS:8123 also fails.
Then, start your HA in safe mode as I already suggested and see if this occurs.
I edited the post above, HTTP with local address works this time. DuckDNS logs seem fine. Let me try safe mode.
As you might know: it’s most likely either a problem connecting remote or local.
Establishing in which cases your issue occurs is necessary to solve it.
It’s getting stranger and stranger. I can locally connect from iPad using ip with http but not from iPhone with same WiFi network, same browser.
Edge says:
Hmmm…diese Seite ist leider nicht erreichbar
Die Verbindung scheint 192.168.178.47 geschlossen zu haben.
Versuchen Sie Folgendes:
- Die Verbindung wird überprüft…
ERR_CONNECTION_CLOSED
Try without this
Deactivated it and restarted. No effect. Only edge now connects locally in iPhone and safari on iPad. Other local tries show connection closed see above. HTTPS not working.
Are you using vlans in your network?
No VLAN.
All other network devices work without problems. It’s only my Home Assistant having trouble.
Probably not going to help but you can try changing the ethernet cable and using a different port.
Ok let me try that.
I have a switch connected to my FritzBox router. Before I had the switch my Home Assistant went offline once restarted Router and Raspberry. Needed to change the plugs in the ETH ports randomly and it worked again.
And you don’t have a second router right, only this Fritzbox is handing out IPs and wifi?