Fix Rocky 9 Gateway Timeouts

My Rocky Linux 9 server, running Nginx as a reverse proxy in front of Apache and managed through CWP, has suddenly started throwing 504 Gateway Timeouts across every hosted site. CWP itself is stable, so I’m focusing on the underlying stack.

The symptoms point to memory-related limits—possibly in the kernel, Nginx worker processes, or Apache’s prefork settings—starving requests until they fail. I need someone who can jump in over SSH, confirm (or rule out) the memory configuration theory, and then tune whichever layer is responsible so pages load instantly again.

Here’s what I expect:
• Diagnose the exact bottleneck (RAM limits, swap, Nginx proxy buffers, Apache MaxRequestWorkers, etc.).
• Apply the safest performance tweaks without breaking CWP’s integrations.
• Document every change so I can repeat it after future updates.
• Prove the fix by loading multiple sites concurrently—no more 504s.

You’ll be working with Rocky 9, CWP, Nginx, and Apache; familiarity with systemd, top/htop, vmstat, and log analysis tools is essential. I’m ready to grant root access as soon as we agree on a timeline.