just a quick question - is it wise/possible to mix phoenix websockets support into existing rails application? Rails websocketa are ‘slow’ when compared with elixir/phoenix, however rewriting whole app is not possible. That’s why idea of just adding websockets via elixir/phoenix.
What do you think?
It is definitely a viable approach. How are you planning to communicate between the elixir and rails apps?
I did this with rails 4 and instead of phoenix used node.js with socket.io and redis pub/sub. I used to push events to redis from the rails app and the node.js app would just read these events and send it to the connected sockets.
Initial idea was to use redis or rabbitmq and then perhaps elixir/phoenix to manage websockets.
Was it complicated in your case?
Mine was a simple use case of streaming the logs of a CI system. So it was very straight forward. Redis/RabittMQ should do the job. You’ll also need to think of a strategy to do authentication if your websocket connections need them. Overall, it shouldn’t be very difficult.
It may also be possible for your rails code to act as a Phoenix channel client and send messages to Phoenix which would then broadcast it as necessary removing the need for redis or rabbitmq. However, I am not sure if that is a good strategy.