How to run multiple rasa or rasa x bots on one server

Hey @erohmensing , I have similar Question, just to confirm from your side that Mulitple bots/domains is not possible in single Rasa server Setup ? can you please confirm this also for latest version of Rasa.

Thanks

Multiple domains is supported (a bot can handle many skills in one), but multiple independent bots are not supported on one server.

Is it possible to deploy multiple bots using single instance? Like One for Normal Q and A’s, another for like different purpose. I mean to say that like I have already deployed a bot using Azure instance (Docker-Compose installation) now i want to deploy another bot in the same Azure instance but for different purpose.

@erohmensing can you reply on this question?

Is it possible to deploy multiple bots using single instance? Like One for Normal Q and A’s, another for like different purpose. I mean to say that like I have already deployed a bot using Azure instance (Docker-Compose installation) now i want to deploy another bot in the same Azure instance but for different purpose.

@sssaha143 Yes you can host multiple Rasa Open source on one Compute Instance using different separate open ports for Rasa NLU and Action Servers.

For ex : Rasa Bot 1 NLU : localhost:5005 Rasa Bot 1 Action server : localhost:5055

Rasa Bot 2 NLU : localhost:5006 Rasa Bot 2 Action server : localhost:5056

and so on …

Assuming the ports used are open and unused.

@athenasaurav , by doing that infrastructure cost will be increases and one bot consumes 1.5GB of RAM because of tensorflow library and if I have 8 bots, 8GB RAM will be consumed by them, in future if I have more bots, then I need to increase the RAM size which will increase infrastructure cost. So, do you know any other alternative to run multiple bots under single rasa instance?

No @bharath-madduri. I haven’t been able to figure out a solution to this.