TRITON-2516 optimize cpu_detailed cmon gz plugin

[jsadowyj]

Overview

This PR modifies the cpu_detailed plugin to use awk efficiently. Some rudimentary testing shows a 98% speed gain after this change.

AI Disclosure

@jperkin worked with Claude to optimize this script.

Problem Statement

The average time to execute the cpu_detailed plugin on one of our EPYC CNs crosses well over the default three second threshold.

Testing

The version of this plugin released in v0.0.6 was manually replaced with this version on one of the problematic EPYC CNs. The log shows no errors:

[root@a12 (us-central-1a) ~]# bunyan -l warn $(svcs -L cmon-agent)

[root@a12 (us-central-1a) ~]# echo $?
0

And the execution time decreased sharply according to our CMON metrics:

Additional Notes

IIf anyone has some suggestions on how to optimize the runtime for the plugin, I am open to suggestions. I would much rather optimize the plugin than increase the timeout.

I already refactored it to use associative arrays in #2, which greatly improved performance, but perhaps there is something else I have not considered.

Thank you @jperkin for the optimization!

[danmcd]

When you get far enough along, make sure your test results are in the ticket.

[danmcd]

Thanks for the testing notes in the ticket. Before I approve this, a couple of quick questions:

1.) The single-awk-run was Claude-assisted? If so, the assisted-by line needs to go into the commit.

2.) But it was suggested by @jperkin ? If so, a Portions-by-jperkin line needs to go into the comment as well.

[danmcd]

Questions answered offline, and will be reflected in the squash-and-merge commit message.