In a situation where you have a huge amount of data/users, you can get away with not running it to reduce the traffic of presence processing. If you are not running it, the user's online/offline status and notification will not work properly.
Set the following environment variables to enable Prometheus metrics:
Enable Prometheus metrics endpoint
Port of Prometheus metrics endpoint
Once all services are up and running the web socket connections should be targeted to ddp-streamer containers, the configuration depends on the reverse proxy you have set up, but you need to change the following routes:
If using Kubernetes, an Ingress like the following can be used:
- host: your-hostname.rocket.chat
To summarize it:
You just need to deploy the reverse proxy to split the communication.
Run all the above services pointing to the NATS and the MongoDB.
Run Rocket.Chat according to the above-mentioned variables.