How can we use streaming capabilities in a distributed workload

[paulRbr]

Pre-submission Checklist

• I have checked that this question would not be more appropriate as an issue in a specific repository
• I have searched existing discussions and documentation for answers

Question Category

• Protocol Specification
• SDK Usage
• Server Implementation
• General Implementation
• Documentation
• Other

Your Question

Hi,

Thanks for adding the stateless attribute to the SSE transport layer of the lib. However, I have a question on how do you use this lib to be able to do streaming responses and actually keep sessions in a production-like environment (I.e. more than 2 workers)?

Do you think the streamable http transport layer could be configurable to be able to receive any kind of “store” for the sessions (I'm thinking of a Redis based store) in place of the current memory hash? I'm not an expert in transport layers, so I'm not even sure we can do streamable http with distributed streams across web servers.

If you folks have any pointers on that, I'd be happy to help and contribute a PR to the lib if possible 🙏

Thanks for the help!

P.s.: sorry for the duplication I first posted my question on the stateless issue here. Let me know if I should remove one (the issue comment) or another (this discussion) 🙇

P.s.2: there seems to be a similar demand in the python-sdk