Hello folks. Got an devops related issue for an EE site that I need to know if I'm on the right trac
How Do I?
Hello folks. Got an devops related issue for an EE site that I need to know if I’m on the right track to addressing.
Site is EE 7.4.x. PHP is 7.4 (yes, I know – in process of getting this addressed) The site has been going down - not crashing the server completely, but the frontend becomes unreachable, CF reports as “server down” but I can still SSH in. On DigitalOcean’s traffic graphs, the CPU is red-lined at 99.5-100%.
Of note – there is not a corresponding surge in traffic on CF or the DO graphs.
I didn’t happen to run ‘top’ before I implemented “under attack” mode yesterday, which made the CPU return to normal.
I’ve been running it since then, though.
Does the site have any really heavy query loops (channel:entries, etc) that need to be cached?
I watch php-fm7.4 occasionally spike to like 50% (a couple of them), then thereafter, mysql surges beyond 100%
It does do template caching on some things
since it’s on 7.4 atm we can’t install something like SPeedy, tho that is the plan once it’s on 8.2
My suspicion was that something in either PHP or mysql process was getting stuck to make the CPU redline and stay that way for more than just moments
What about storage?
disk? it’s at like 50%
Good to rule that out
it was almost 90% before but that is because of some image directories that were replicated unnecessarily. I got rid of those and it’s still having this problem
Any massive images in the user uploads? I’ve seen a 40mb+ image slow a site to a crawl because the transforms were taking a very long time.
We do have JCOGs installed to combat that, tho it’s set to not work (e.g. render the image) unless they keep it under 4MB now.
they absolutely did have big ones before but we advised them about that to knock it off. LOL in kinder words of course haha
i gotta run to an appt but i’ll check back later if there are more questions or suggestions for me to try.