I’m not too familiar with R2DBC journal plugin but is this required if I want to use eventsBySlice in Event projections? Can I use this plugin on either a Cassandra or Dynamo Db backend?
Supported databases by the Akka R2DBC Persistence plugin right now are Postgres, Yugabyte and H2. Other Postgres protocol and query language compatible databases could work, but it is not anything we test.
Both Cassandra and Dynamo are quite different beasts, so those will not work with the R2DBC plugin.
The Akka Persistence Cassandra plugin does not currently have support for the events by slices query.
The Dynamo DB persistence plugin is community maintained, so I’m not sure what the status is there.
Thanks Johan. I take it R2DBC is more of a JDBC replacement then. Is there no need for slices with a Cassandra and/or DynamoDb journal for event projections? I think that eventsByTag is a bit cumbersome when you want to scale out your projections. The way I can think of is to introduce new tags for the new scaled out projections and shut down the old projections when all the old event tags are consumed.
Yes, the capability to change the number of partitions after the fact is the main benefit with events by slices. The brokerless eventing used for projections and Replicated Event Sourcing in Akka Projection gRPC depends on it, and is not possible to use with events by tag. Events by slices would be interesting for other plugins as well, but it is not something that has been worked on.