I set up a fresh OH3 based on a docker install mainly following the guide from Andreas Laub.
With the installation of the network binding and about 10 network item the CPU consumption is raising approximately all 1,5 hours to a relatively high level. Please see attached image. The measurements are inside the container. Nevertheless I have no idea why and how the network binding would cause this behaviour especially all 1,5 hours.
It really started this behaviour with the items activation and was stopped immediately with the items deactivated on OH3, as you see by the again flat line after the deactivation.
You graph “system cpu load” which means that any of tools you run on your pi might cause this spike. If you ingest larger amount of data into influx at frequent base it can simply do compaction of storage files which takes CPU and bunch of I/O at the same time.
Thanks for your thoughts about that. That sounds plausible. I’ve activated that network binding again with no Mobile (the talkative ones) and the with one Mobile.
There is not effect like the above visible. Tomorrow I’ll try to activate the other mobiles and see if the behaviour can be reproduced.
I doubts that it is four mobiles alone, but together with many temperature and movement sensors, that are also very talkative. Plus maybe the limitation of ressources that are bound to the container?
So maybe I just hit a certain threshold… We’ll see…