Two parallel applications on same machine -> two separate streams or single stream & zmq?

[am11y]

I have two different, parallel applications which both need Kline and User data stream on the same machine. Do you suggest I use:

1. Single subscriptions and distribute data between apps using zmq or similar?
2. Use subscriptions for each of the apps - make duplicate streams?

[oliver-zehentleitner]

I can't give you a general answer to that!

If it's Klines for 1-2 markets, I would do it per process. If you are monitoring 1000 markets, I would do it centrally and outsource it to an extra process. Maybe even two of them, then you can restart a receiver process and the other process continues to work in the meantime.

How time-critical is your bot? How do you deal with gaps in the data, is that a problem for your bot?