Play joining Akka Cluster

I have a quick question regarding what is meant by embedding play-akka-http application in an Akka cluster. Does this mean there is a replica of play on each node or is this a play singleton participating in the cluster?

Additionally, is it a good idea (or even achievable) to put a standalone play instance outside the cluster which can then connect to the akka cluster in the same vpc / subnet (play can run its own classic actor system). I assume in this case Play won’t have any injected references to actor systems in the cluster so what would be the best way to achieve that?


I can’t answer the first part, the wording leaves it open for multiple interpretations.

On the second part : That is usually done using Cluster Client or its successor, gRPC. That gives you the option to route a message inside the cluster, assuming you know the target address.

1 Like

Hi @farooqkamal and @igmar,

please do no use cluster client to communicate two processes. The best way forward is to expose gRPC endpoints on your Akka Cluster and consume them from the Play application.


1 Like

gRPC isn’t always an option. But I agree, it should be a first preference.