Abrupt steep ascent of processor use and temperature

The problem

I can observe steep ascents of processor use, processor temperature and CPU processor. CPU and processor temperature rise from about 33°C up to 65°C, and processor use from 2% up to 71%. Moreover, swap is growing - very slow, but - constantly (from 0% to now 18% with memory use around 36.5% (= 1.6 GB of 4 GB).

I have a rather basic setup. And except for some automations and sensors, I have excluded everything in the recorder (excluding the domain light (18 smart devices) for example lowered the memory by 4%).

I disabled all automations and rebooted, but there was no difference. Following advice in the forum, I installed the Profiler integration and attached the files below: Archiv.zip

ha info:

arch: amd64
channel: stable
docker: 20.10.14
features:
- reboot
- shutdown
- services
- network
- hostname
- timedate
- os_agent
- resolved
hassos: null
homeassistant: 2022.5.1
hostname: xxxx
logging: info
machine: qemux86-64
operating_system: Debian GNU/Linux 11 (bullseye)
state: running
supervisor: 2022.05.0
supported: true
supported_arch:
- amd64
- i386
timezone: Europe/Berlin

What version of Home Assistant Core has the issue?

core-2022.5.1

What was the last working version of Home Assistant Core?

No response

What type of installation are you running?

Home Assistant Supervised

Integration causing the issue

No response

Link to integration documentation on our website

No response

Diagnostics information

No response

Example YAML snippet

No response

Anything in the logs that might be useful for us?

No response

Additional information

No response

[bdraco]

It looks like the cpu usage is the stream integration taking snapshots from the foscam integration

@bdraco Just saw this old thread…just wanted to clarify that foscam does not use stream to generate snapshots (it looks like snap_picture_2 just grabs a pic from the camera using http). Also, it seems like the py-spy output for threads can be misleading as it does not reflect that the thread is often idle with the GIL released - the http operation for the snapshot as well as the next operation in stream should be outside of the GIL for most of the time.

I don’t think stream is the cause of this one. I expect this one is outside of core and in a custom component somewhere that has yet to be identified

[bdraco]

oops! I mean re-enabling the sensor 👍