Investigating Load/Resource Usage #

I’ve honestly never used any of the dedrads given to analyze resource utilization. I actually only recently became aware of some of these.

• check_user
• check_server
• check_apache
• fpm_status

These are just wrappers for other utilities that I would argue are part of the fundamentals we should be teaching as part of the APS pipeline.

Things like ps, sar, top, htop, free, uptime, apachectl, mysqladmin, and reviewing both the Apache and PHP-FPM logs for issues should be the focus here.

exim and/or eximstats might be appropriate here as well.

We can introduce our wrappers outlined above after we’ve taught the fundamentals.

The nlp script, although it’s a bit broken is also helpful for understanding resource usage spikes related to specific sites. Introducing the Apache domlogs is probably a good thing to cover here.

For resource utilization issues on shared, I’ve actually used the check_software script more often than any of the scripts provided above. If anything that script might be a candidate for introduction here, but probably more appropriate in the T1E training, if I’m honest.

I would argue that nlp and check_software are better candidates for introduction here as opposed to the four dedrads that are actually introduced, if we’re going to introduce dedrads in the pipeline at all. I really think we need to make the pipeline dedrads and cPanel scripts agnostic. Linux and sysadmin fundamentals that can be more or less universally applied skills should be the focus. We want agents who know how to reason through problems rather than just use the tools we’ve built. Our dedrads or cPanel scripts can become a crutch otherwise.

This section introduces GTMetrix as a tool to analyze performance issues. That’s a useful tool, but again the official T1E training should probably introduce this. I feel like competency in diagnosing performance issues on websites even if it’s just to advise a developer is something a competent T1E should already know how to do.

APS is where we need to learn how to diagnose, troubleshoot, and mitigate performance issues on the server level.