I have a very annoying issue with my HA setup. HA currently sits on an rPi4 on my network, cabled into the core of my LAN (direct access to wifi AP and to broadband firewall), which has sufficient memory free and sufficient CPU free at all times for me not to be concerned about performance. I have a significant number of entities (I’m not sure how to get a number of them but its over 1000) and a lot going on, plus a lot of HACS additions and other resources and dashboards, but by and large it runs happily with no problems. Its up to date.
I have three ways I can access it (same as everyone else does) - homeassistant.local, IP address, or proper FQDN (a duckdns address). It runs as HTTPS as I have it accessible externally (and I need to put nginx in to take that load off the core system as external access is extremely rare for me, but thats another day’s job). Of minor interest is that calls to the FQDN actually resolve to the external IP address, and traffic from internal hosts hairpins via my router and back to the pi, but as per below that fact is not a cause of the issues.
The problem is that intermittently I get endpoints (tablets, phones, laptops - I have them running all OSs and of all types!) randomly bring up either the status popup at the bottom saying connection lost, reloading, or the ‘front’ loading HA logo with ‘Unable to connect to Home Assistant’ and the countdown timer. It will happen on one browser or app on one device whilst I am happily still browsing on HA from another uninterrupted - by way of convenient example my tablet is beside me looping around the 59 second reload counter while I’m watching the logs and hunting around HA on my mac to see if I can see anything to add to this post.
Without fail, if I change the address being used in my browser (or HA app too) to a different one of those three options above, it will load fine. It fails equally on all three addresses too, so its not any single one causing the issues (its not SSL certificate erros, for example). I just slowly circulate round the three addresses across all my devices as this happens.
I can’t see anything in the logs either to show there had been an issue.
Has anyone got any idea whats happening?
–Chris