Ensuring Atomicity and Scalability in Redis Backplane?

[locus84]

How does Redis Backplane ensure the atomicity of room state?
It seems like there could be an issue where the group state becomes out of sync if, for example, saving the group state in Redis succeeds but publishing fails.
The only solutions I can think of are using a distributed lock or executing all redis commands at once with Lua.

Additionally, Redis Cluster supports sharding but stackexchange.redis does not support spublish/ssubscribe, scaling out seems difficult.
Do you have any other recommended approaches?(searched about kafka or evony proxy for redis)

[locus84]

I see Nats jetstream can be viable solution and there's beta support for nats in multicaster package, fantastic!!