[Epic] SignalR Redis improvements

[BrennanConroy]

• Figure out if Clustering works or what it would take to make it work - It works
• If the redis connection is lost what should the behavior be
  • Call ConenctAsync again for a new connection? Redis reconnect problem #5393
  • Buffer Client.X.SendAsync calls instead of throw?
• Investigate new streaming APIs Investigate redis streams for the backplane implementation #5313

[ghost]

Thanks for contacting us.
We're moving this issue to the Next sprint planning milestone for future evaluation / consideration. We will evaluate the request when we are planning the work for the next milestone. To learn more about what to expect next and how this issue will be handled you can read more about our triage process here.

[ghost]

We've moved this issue to the Backlog milestone. This means that it is not going to be worked on for the coming release. We will reassess the backlog following the current release and consider this item at that time. To learn more about our issue management process and to have better expectation regarding different types of issues you can read our Triage Process.

[sikri-eic]

Figure out if Clustering works or what it would take to make it work

Our experience shows that Redis clustering (with Azure Redis Cache) does not work (of course, it's possible that we configured something wrong). So, we had to move our SignalR app to a non-clustered Redis cache apart from all the others using the Redis cluster (which is not a great way to scale and is not cost-effective). So, I'm very interested in learning what your findings (and solution) on this will be. Subscribing...

[ghost]

Thanks for contacting us.

We're moving this issue to the .NET 8 Planning milestone for future evaluation / consideration. We would like to keep this around to collect more feedback, which can help us with prioritizing this work. We will re-evaluate this issue, during our next planning meeting(s).
If we later determine, that the issue has no community involvement, or it's very rare and low-impact issue, we will close it - so that the team can focus on more important and high impact issues.
To learn more about what to expect next and how this issue will be handled you can read more about our triage process here.

[BrennanConroy]

Tested Redis Clustering with Azure Redis Cache. No additional configuration or code changes were needed on the server, just the connection string like usual.

Tested with 2 servers on separate shards, one client on each server.
Tested that broadcast to all clients works, and that sending to a specific connection works (this hits the "group" subscription code path).
Also, forced 1 shard (primary and replica) to restart, server automatically reconnected and maintained it's subscriptions, tested via broadcast and send to specific client again.

[sikri-eic]

@BrennanConroy Thank you for testing this. We could not get Redis clustering to work on Azure Cache for Redis. I cannot remember what sort of errors/problems we got.

No additional configuration or code changes were needed on the server, just the connection string like usual.

This is interesting. Our connection string is in the following format as of now: <url>:<port>,password=<password>,ssl=True,abortConnect=False

I wonder if this has anything to do with the problems we have experienced.

Anyway, I will retest this as soon as I can.

[BrennanConroy]

Our connection string is in the following format as of now: <url>:<port>,password=<password>,ssl=True,abortConnect=False

That's a normal connection string.

[sikri-eic]

That's what I thought too... Will test as soon as I can. Thanks.

[tomap]

Hello,

My understanding was that the commands used for redis are publish and subscribe which are not spread across nodes of a redis cluster.
So I'm not sure to understand how the current implementation of backplane on redis is compatible with redis clusters.

Redis 7 introduced ssubscribe and spublish which are compatible with redis clusters. (Sharded)

But maybe there is something I did not understand in either redis clusters or SignalR backplane?

[BrennanConroy]

I think this is the key paragraph:

Sharded Pub/Sub helps to scale the usage of Pub/Sub in cluster mode. It restricts the propagation of message to be within the shard of a cluster. Hence, the amount of data passing through the cluster bus is limited in comparison to global Pub/Sub where each message propagates to each node in the cluster. This allows users to horizontally scale the Pub/Sub usage by adding more shards.

So these new Redis commands help performance, but that doesn't mean Sharding will not work with SignalR today.