Traccar with Kubernetes
Traccar doesn't support horizontal scaling. It has internal cache.
So, Do you have any recommandation for spectification of VM?
For instance,
1000 Device, 100 Users (for web admin that used the Websocket also).
4vCPU / 8GB Ram?
It all depends on the message volume.
You means like the amount of device and update frequency right?
Yes.
Got it, How about 1000 Device with interval update is 15 second
and 50 Users (for web admin that used the Websocket also).
That's pretty high load. You should probably do your own load testing to figure out what works.
Sorry to get into this old topic.
is it possible to use only the api with kubernates?
I don't use the web interface.
Why not? Nobody is going to force you to use the web app.
Hi, Anton Tananaev
I would like to ask some question about best practice of Traccar as container.
As I know the Traccar docs have the docker container example.
Then If I run the container on the Kubernetes cluster. It's possible ? (Mean the Traccar is support for Kubernetes)
The thing that I want to implement It's kind of stateless services.
So I read this topic https://www.traccar.org/forums/topic/limitation-of-server-for-scalability/
It's doesn't any mention to the stateless or Kubernetes architecture. Just only the host the Traccar on VM. And the problem of database connection.
It's would be great If have any example case for Traccar with Kubernetes.
Please consider the image flow below
The question is It's possible to do like that?
and I have to consider anything before implementing? or Traccar limitation? and compatiblity of Traccar and kubernetes?