review: added comment for config key "subscriber-pool-size"

Signed-off-by: Thomas Jaeckle <thomas.jaeckle@bosch.io>

internal/utils/ddata/src/main/resources/reference.conf

@@ -57,7 +57,13 @@ ditto {
     number-of-shards = 5
     number-of-shards = ${?DITTO_DDATA_NUMBER_OF_SHARDS}
 
-    subscriber-pool-size = 1
+    # The number of subscribers in each subscriber pool.
+    #  Akka deserializes (remote) messages addressed to 1 actor sequentially, which makes deserialization throughput
+    #  unable to scale with the number of CPUs given to the instance.
+    #  Each instance starts a fixed number of `subscriber-pool-size` subscribers.
+    #  Published signals are sent to each subscriber according to the entity ID.
+    #  This allows signals to different things to be deserialized in parallel.
+    ^ = 1
     subscriber-pool-size = ${?akka.remote.artery.advanced.inbound-lanes}
     subscriber-pool-size = ${?REMOTE_INBOUND_LANES}
     subscriber-pool-size = ${?DITTO_DDATA_SUBSCRIBER_POOL_SIZE}

internal/utils/pubsub/src/main/java/org/eclipse/ditto/internal/utils/pubsub/actors/AbstractSubscriber.java

@@ -48,6 +48,7 @@
 abstract class AbstractSubscriber<T extends SignalWithEntityId<?>> extends AbstractActorWithTimers {
 
     final DittoDiagnosticLoggingAdapter logger = DittoLoggerFactory.getDiagnosticLoggingAdapter(this);
+
     final Class<T> messageClass;
     final PubSubTopicExtractor<T> topicExtractor;
     final AckExtractor<T> ackExtractor;