Avoid failing when the CPU period is zero

[felipecrs]

And instead, fallback to counting CPUs from /proc/cpuinfo instead.

This situation is theoretically not possible, but it has happened as per:

• [Support]: tempio: command not found blakeblackshear/frigate#11755 (comment)
• Fix case where user's cgroup says it has 0 cpu cores blakeblackshear/frigate#13271

[felipecrs]

@christianonpele, I'm trying to understand the situation you described in:

• [Support]: tempio: command not found blakeblackshear/frigate#11755 (comment)

Could you please run the following command within Frigate's container and paste the output here?

# example to get access to Frigate's container
$ docker exec -it frigate-container-name bash

$ curl -fsSL https://raw.githubusercontent.com/felipecrs/cgroup-scripts/78da74851849afa821488131a52e650cc6c6160c/get_cpus.sh | DEBUG=true sh

This would help me understand the issue better, and perhaps writing a better solution, as opposed to blindly add a protection for division by zero. Specially given this situation is theoretically not possible.

[felipecrs]

This was not tested against systems with cgroup v1, but it should work.