If I increase the CPU and memory of the VM where the bridge is installed, can I avoid server load high?

Sometimes the video comes out as black screen.
So I checked the structure of our server to prevent high load.

I checked, and the servers are divided in the old fashioned way.

Our server consisted of a total of 4 types of VMs.

  1. Red box: VM with UI
  2. Green box: VM with gateway running
  3. Blue box: Bridge
  4. Other servers (not pictured): VM where the API that puts data into our DB is running

There were a total of 4 bridge servers.
(In the case of the bridge server, each was contained in a total of 4 VMs.)

In the case of jicofo, the previous developer directly modified the java source code.
Probably calling our api from jicofo.

I couldn’t touch this structure at all because it was so complicated.
Can high load be solved by simply increasing the CPU and memory specifications of the server?

Or should I increase the number of bridges?
(Bridge currently have 4 VMs, should I increase this to 6?)

My company told me to prepare for more users in the future.
So, he asked me to review server expansion.
But I’m not sure if this means to increase the vm or the server’s specifications.

Where do you see the high load and how do you measure it?

1 Like

When doing a video conference, there was a problem that a black screen often appeared.
Then the existing developer could not solve this problem and left the company.
Then suddenly the jitsi project was assigned to me. But i don’t know what jitsi is.

My company told me this.
‘Look at the vm and specs now and prepare for expansion’

I don’t know what this means. So I posted this question.
Perhaps the my company thinks the cause of the black screen is high load.

This may be another issue related with the client-side. First you need to be sure about the cause. If it is not related with the server resources, you cannot solve it by increasing the resources on the server-side.

1 Like

First of all you should start to monitor your jitsi infrastructure components: CPU load, network usage and jitsi connected participants.

Then you should determine in which cases you see your video issues : on all conferences, only when the number of participants increase…

There is a lot mechanism in the jitsi infrastructure that can reduce the number of video displayed in a conference to reduce the load.
The capacity of your jitsi infrastructure really depends on your mean conference size and the number of video stream that are displayed by clients.


1 Like

Black screen doesn’t work in my area.
(Everything works fine on my computer.)

However, when people in other regions join the video conference, there seems to be a black screen.
In particular, people who have a military base nearby often experience a black screen.

But when I went in, that never happened.
I don’t know how to solve this. It is not reproducible to me.
So the previous developer also quit the company.

My company told me to prepare for more users in the future.
So, he asked me to review server expansion.
But I’m not sure if this means to increase the vm or the server’s specifications.

Sounds like they’re behind a restricted firewall and can’t process UDP traffic. You’ll need to deploy a TURN server.


The black screen doesn’t happen all the time, it happens irregularly.
Strangely, when I rebooted the bridge server, the black screen didn’t occur for a while.
So, we are rebooting the bridge server every week.

Currently, each of our bridge VM has (16Core of CPU and 16G of Memory) * 4,
and it is said that it can handle 100 * 4 concurrent users.
(I heard that it can handle a total of 400 concurrent users)

I’m trying to figure out various causes through this site.

It is difficult to solve the issue if you don’t reproduce it.

1 Like