Server Optimization Issue Around Mob Cap

Hello Everyone,

There's a server I've been trying to help figure out lag spike issues for and while we have figured out the source of the lag we've indirectly created an issue with mob spawning causing mob farms to struggle to produce when we have 15-20 players on.

Here's the hardware information followed by some of the key settings I believe are causing this issue.

Server:
- Jar - Fabric v0.17.3
- MC Version - 1.21.10
- Host - Apex Hosting EX Series w/ 16GB Ram
- - Spark says the CPU is described as a Ryzen 9 7950X
- - Apex on the site themselves don't describe the CPU in use other than their plans/sales page where they describe the CPU as a Ryzen 9 5900X in the US (where our server is located)

Mods: Performance mods such as Lithium

Server Config Values:
- Simulation distance: Moved from 7 -> 5 (this appears to be the change that resolved the lag as we saw most of the CPU usage being taken up by mob entities using Spark - Makes sense less chunks being processed)
- View Distance: 8
-Entity-Broadcast-Range-Percentage: 37

Since making this change mob farms seem to be producing at a fraction of what they were and in some moments aren't producing at all.

Utilizing Carpet mod logging on mob caps I can see that the mob cap is nearly full constantly. This makes sense to me since the lazy chunks outside of the simulation distance still contain mobs which count towards the global mob cap but aren't being simulated so they aren't being picked up by the despawn checks.

So my question(s) based on this information is as follows:

- Is there anyway to consistently despawn these entities outside of the simulation distance utilizing Fabric, a performance improving mode or server settings? The goal is to try and stay as true to vanilla as we can.

- If there isn't a way to do this, are there suggestions for moving simulation distance back up to 7 and improving performance of hostile mob entities so they don't eat up the server bandwidth.