Hi, A quick few questions regarding cluster sharding:
Firstly, does sharding shut down actors after a period of time? In the example there's a SetReceiveTimeout call that doesn't appear to be handled anywhere. If it's not automatic, how would I achieve this as I've a process that could last for days as it something that happens off-line.
Secondly, is there a way to only create an actor when the shard region receives a specific message , there are other messages that will be sent subsequent to this first message, but they should only be handled if the actor has been created. Also, how does cluster sharding generate a unique path within each shard?
Finally, taking the customers example from the docs, how would you say query for all customers that have purchased a specific item using sharding? Should separate query actors be created?
Help apreciated as I'm trying to work out whether to use raw persistence or sharding.