Add cpu/cpuset cgroup and procfs data gathering

[arunramani]

Description

Add different ways to collect CPU availability information. This PR gets information from cpu/cpuset cgroups and from /proc/cpuinfo. Also adds 2 more cgroup monitors to emit this info.

The motivation here is to improve the way we figure out available processors since Java's availableProcessors implementation doesn't account for fractional quotas, effective CPU sets and shares. This becomes especially important in containerized setups like Kubernetes. Eventually we will extend RuntimeInfo to be smarter and when we move to Java 11, we can swap to theMetrics class.

Key changed/added classes in this PR

• Added Cpu and CpuSet to java.util.metrics.cgroups
• Added ProcFsUtil for procfs info
• Added CgroupCpuMonitor and CgroupCpuSetMonitor

This PR has:

• been self-reviewed.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.

[abhishekagarwal87]

LGTM

[clintropolis]

lgtm after CI 👍

we should probably add docs to https://github.com/apache/druid/blob/master/docs/configuration/index.md#enabling-metrics and maybe to https://github.com/apache/druid/blob/master/docs/operations/metrics.md though the latter does not have the other cgroups metrics. im ok if docs are added in a follow-up PR

[abhishekagarwal87]

@arunramani should be good to go once we eliminate the possibility of division by zero error.

[arunramani]

@arunramani should be good to go once we eliminate the possibility of division by zero error.

It should just return Infinity without exceptions but I've set it to 0 since I have no idea how monitors react to infinity values.