I noticed that when HP23 has been running for a few days, the average CPU utilisation creeps up to 35% in idle mode, with periodic spikes up to 85% which seems quite high for the new hardware. When I reboot, this drops to a much more reasonable 5-10% idle with spikes up to 60% and it stays like that for a while (see screenshot). So something seems to be consuming 25% CPU continuously.
Reboot at yellow line:
I have about 60 Zigbee devices, mainly Aqara sensors, generic bulbs and Tuya Plugs. I’m using a limited number of advanced flows (about 20), and none of them very complex. I also tried to rule out flows and Apps but with no luck. Checked Insights for each App CPU + Mem utilisation, and nothing really stood out. Worth to note that there does not seem to be a direct relation with how busy the network is, I see the same pattern during the night.
Before I open a case with Athom: does anyone else recognise this behavior?
BTW: Would also like to see the periodic peaks explained, but that’s for another thread.
Is there any way we can get detailed (debug) logging as end users?
[Update] After 4 days of a low baseline, last night around 23:30 the average CPU utilisation suddenly went up with 25% again Rebooted Homey to solve it.