The Nextcloud Hub 9 docker load is nearly 100% all the time. I am using a contabo VPS 1 SSD Disk Space: 400 GB CPU cores: 4 RAM: 6 GB. TOP is showing 5.6% where Nextcloud>Administration Settings>System is around 16% now (for the first day it was nearly 100%). There are no active users yet. In comparison my live snap version of Nextcloud version 20.0.14 has very low load of around 0.15-0.23, while allocated much less memory 3.85 GB, and is only a dual core system with 15 active users. What I notice the most, other than the high load, is that the swap usage on the Nextcloud docker instance is 0. Please advise if there is anyway to lower the load. It looks as if SWAP isn’t recommended now based on some of the other answers here.
Nextcloud snap 30.0.5snap1, SWAP disabled, 4 CPU, 6GB, LXC/LXD, 5 Users see SCUBA’s Nextcloud snap server setup & specs, sync cronjob every 4H.
newer snap Hub 9 30.0.5 for reference? 
Some of the metrics you seem to be using are apples to oranges so it’s challenging to draw any conclusions from them one way or another. Also, more granular data would be needed anyhow to draw any formal conclusions.
Rather than playing a guessing game:
- What processes specifically are consuming CPU?
- Which “Docker” image/stack of Nextcloud are you using? There are many and there is a massive difference between how they’re set up by default and what components of the stack they come with.
- What apps are installed in Nextcloud and what is your configuration?
Couple other thoughts:
- A load average of 0.15-0.23 on a dual core system = 7.5% to 12.5% CPU usage
- A VPS is an inherently shared resource and numerous factors play into whether you actually have access to the CPU/RAM/Storage specs that are printed on the tin. This is, in large part, why there is a huge variance in performance across providers that appear to (superficially) have the same specs.
- It can be challenging to compare a VPS to other environments – particularly without knowing more details about the underlying platform, it’s configuration, its (over) provisioning policies, etc.
While we’re making arbitrary comparisons: I have, here in my office, 28 (!) separate Nextcloud stacks running in Docker on an Ubuntu Server and my load average is 0.03/0.19/0.23 with 4988/2641 MiB (Mem/Swap) used (total not per deployment). All that really tells you is that some combination can result in that resource usage. You’d have to dig deeper to know whether that has any relevance to your use case.