0.8.11

What's Changed

• Add actor watching functionality, actors can now watch other non-child actors.
• Unify error handling for non-registered remote messages sent via send & notify by @RedKinda in #25
• Feature/exact protobuf version by @AdriaanPrinsloo, fixes issue #32 in #34

New Contributors

• @AdriaanPrinsloo made their first contribution in #21

Full Changelog: coerce-0.8.10...coerce-0.8.11

0.8.10

** Coerce no longer requires tracing_unstable + tracing/valuable features **

• Fixes sharding inconsistency where a shard would sometimes continue running if re-allocated on another node after losing a coordinator
• Minor optimisation in networking sessions - IDs have been changed to sequential i64, rather than generating a UUID per session.

Full Changelog: coerce-0.8.9...coerce-0.8.10

0.8.9

What's Changed

• Receiver can now be cloned, thanks to @AdriaanPrinsloo in #21
• new_actor, send and notify instrumentation to use level=debug, thanks to @cameronbraid in #18

New Contributors

• @AdriaanPrinsloo made their first contribution in #21

Thanks guys!

Full Changelog: coerce-0.8.8...coerce-0.8.9

0.8.8

What's Changed

• feat: propagate current span over notify/send boundary by @cameronbraid in #17
• minor change: actor refs no longer copy the system ID

Full Changelog: 0.8.7...0.8.8

Big thanks to @cameronbraid

0.8.7

• Build fixes when not using "full" feature.
• Update utoipa, fixing dependency build issues
• New persistence APIs: write_message_batch, read_messages, read_message, delete_messages_to
• Minor optimisations and improvements

0.8.5

• Small release, just an example fix within the documentation

Full Changelog: 0.8.4...0.8.5

0.8.4

What's Changed

• Added optional JWT authentication between Coerce cluster nodes
• Started initial crate documentation: https://docs.rs/coerce/latest/coerce/
• Updated utoipa + utoipa-swagger-ui (#11)

And more improvements, see full changelog

New Contributors

• @cameronbraid made their first contribution in #9

Full Changelog: 0.8.3...0.8.4

0.8.3

• Extendable system health check - add the ability to register custom actor implementations to be part of the health check:

If any of the system actors take longer than 1 second to respond to a GetStatus, the health check will report the health status as Degraded.
If an error is returned from any of the actors, the health check will report it as Unhealthy.
Health check also provides useful information such as currently active cluster peers and version information.

• Shards will now start entities asynchronously and are no longer blocked by individual entity recoveries, messages will be buffered until the shard receives confirmation that the actor has started successfully.

• HTTP API Swagger + openapi powered by utoipa crate.

Full Changelog: 0.8.2...0.8.3

0.8.2

• Hierarchal actor path
• Feature flag fixes
• Removed all targets from logging, now it will use the automatically generated module path
• Coerce HTTP API now offers a built-in prometheus metrics exporter
• Cleanup

Full Changelog: 0.8.0...0.8.2

0.8.0

🔈 Announcing Coerce 0.8.0

• Feature flags: "remote", "persistence", "metrics", "sharding", "api"
• ActorSystem memory usage improvements, now it is just a wrapper over an Arc<ActorSystemCore>, allowing for much cheaper cloning.
• RemoteActorSystem can now be cleanly shutdown
• Cluster node termination fixes and improvements, nodes will now emit a message to peers upon system termination and leaders should be re-allocated immediately, which will help improve cluster reliability as a whole
• Multiple persistence backends can now be configured based on the actor's type
• Redis persistence journal provider has been refactored and is now backed by its own actor
• New Kubernetes discovery provider, you can now automatically discover peers hosted in Kubernetes, based on a configurable pod-selection label
• Cluster sharding chat example application now uses protobuf for serialisation throughout
• Updated the cluster sharding chat example application with kubernetes deployment manifests (kustomize)
• Added specific unit test for cluster heartbeat / state consistency after losing the leader node
• Added unit test for shard rebalancing after losing the leader node and more
• Several actor persistence resilience improvements
• Actor describe functionality, ability to examine the actor hierarchy via Describe message or the HTTP API
• Ability to add tags to actors, which can be retrieved via the describe functionality
• Error propagation and handling improvements
• Dependency updates

So much more!

Full Changelog: 0.7.0...0.8.0