Very higher server load

Description

I have a Rocketchat server, installed as a Snap on Ubuntu, that is struggling under load. The machine frequently reaches a load average of 20 or even 30, and at those times node frequently takes up 300% of CPU or more, while using up all its 16Gb of RAM. I’m not noticing any particular errors in the log.

How might I alleviate this? Hardware upgrade? OS tuning? Some kind of proxy? (I tried to use Cloudflare, but it caused new problems.) As a workaround, I’m just restarting the server whenever it gets like this, and that helps for a little while. Happy to pay for someone’s time to fix the problem if that’s required.

Server Setup Information

  • Version of Rocket.Chat Server: 5.3.4
  • Operating System: Ubuntu 22.04 LTS
  • Deployment Method: snap
  • Number of Running Instances: 1
  • DB Replicaset Oplog: enabled
  • NodeJS Version: v14.19.3
  • MongoDB Version: 4.4.15
  • Proxy: none
  • Firewalls involved: none

Any additional Information

Users

Total: 1240
Online: 325
Busy: 4
Away: 42
Offline: 869

Types and Distribution

Connected: 367
Activated Users: 1237
Activated Guests: 0
Deactivated Users: 2
Rocket.Chat App Users: 1

Uploads

Total Uploads: 9648
Total Uploads Size: 21.43 GB

Total Rooms

Rooms: 1120
Channels: 38
Private Groups: 11
Direct Message Rooms: 1071
Discussions: 0
Omnichannel Rooms: 0

Total Messages

Messages: 34222
Threads: 768
Messages in Channels: 10528
Messages in Private Groups: 7398
Messages in Direct Messages: 16296
Messages in Omnichannel: 0

did you solve this problem?

No, not yet. My workaround is just restarting at regular intervals. But it seems like this may be worth trying.