* Add currentEventsBySlices to Persistence TestKit to make it feature comatible with currentEventsByTag, simplifies some usage in downstream projects
* Add DurableStateStoreBySliceQuery to Persistence TestKit, to make it feature comatible with the tagged queries simplifies some usage in downstream projects
* Add optimal clairvoyant algorithm to passivation strategy simulations
* Add natural zipfian distribution trace using text of Moby Dick
* Add some looping access pattern traces from the LIRS paper
* Add CloudPhysics access pattern trace from LIRS2 paper
* Add Wikipedia access pattern trace to passivation simulations
* akka-protobuf-v3 is anyway a dependency from akka-stream
* akka-remote can remain provided by adding lazy on the WrappedPayloadSupport
* tested with a sample with only akka-persistence and akka-persistence-query
* in multi-node tests we create the volume /opt/volumes/media-driver
and run with -Daeron.dir=/opt/volumes/media-driver but when starting
the embedded media driver a unique part is added to the directory name
so it becomes like
Started embedded media driver in directory [/opt/volumes/media-driver-ClusterLogVerboseEnabledSpec-27c2fe72-c2b2-4ef4-ab87-a783f5d9ece4]
* this adds the unique part as a sub-directory instead, so that it's still in in the
/opt/volumes/media-driver/ volume
* I don't know if this will make things more stable but the defined volume wasn't really
used as intended
* wip
* Don't set crossScalaVersions at the toplevel
* scalafmtSbt
* Accidentally commited
* Set default scalaVersion
* Set top-level scala version to default
* Add comment explaining why to +~
* Simplify the variations for passivation strategy settings
* Abstract out double-linked list operations for recency and frequency
* Add optional dynamic aging to least frequently used passivation strategy
* Add segmented least recently used passivation strategy
* Split entity passivation specs into multiple files
* -XX:+AlwaysActAsServerClassMachine in CI
* Remove GHExcludeTest from all tests, since we have included those in CI workflows
* Exclude a few failing tests from CI
* run CI with -Dakka.test.tags.exclude=gh-exclude again
* fails frequently in the aeron-udp multi-node tests
* could be that it requires more resources since it's starting
many actor systems in same jvm
* noticed heartbeat send delay in logs, but not unreachable
* adjusting config to what we have in some other tests
* from logs it looks like the replicated event from the first
incarnation of the actor isn't stored before the second
incarnation is replayed
* stop and wait for termination before starting second incarnation
