Maybe one day a month, something seems to go totally crazy on my home LAN, causing frequent WiFi device disconnects. I have an ESP like device which sends a periodic heartbeat ‘ping’ to a mosquitto server. If the ping isn’t heard, the sensor ‘expires’ and an alert is triggered. Basically for whatever reason the message didn’t make it to the broker.
This isn’t the same as the last will and testament. I have a separate alert for that, which does work, (but hasn’t been triggered at all today).
Anyway I’ve been getting spammed all day by these alerts, UP and DOWN all day long. I noticed the first one I got was at 12:54, which is when my HA log starts showing a load of entries for things taking longer to update than they should. It’s also when my swap usage starts climbing 
Looking at my Unifi access point, the devices in question seem to enjoy WiFi ‘experiences’ of between 95 and 98% so WiFi connection problems don’t seem to be the problem.
Straightforward extremely hard to trace memory leak causing this do you think?
A year ago my swap usage was always at 99% because of a bug in HA but that got fixed a few releases back and it’s been <20% ever since
Thanks