Release v10.9.4

Changed

• Switching from our own implementation of ReminderTable to the one found in Orleans.Provider.MongoDB project.

Release v10.9.3

Changed

• Switching from our own implementation of ReminderTable to the one found in Orleans.Provider.MongoDB project.

Release v10.9.2

No release notes

Release v10.9.1

Summary

We were noticing inconsistencies and wrong results when catching up events. This is now fixed

Added more integration tests around Observers and specifically Catchup for Observers in different states

Added

• Functionality for directly registering Reactors and Reducers by type from the Client

Changed

• Contracts for handling of events for Reducers and Reactors to include Partition key

Fixed

• Catching up observers
• Replaying observers
• An issue where handling events for ar Reducer would fail if the method was an asynchronous Task
• An issue with handling events in parallel across multiple partitions

Release v10.9.0

Summary

Optional summary of the PR here. The GitHub release description is created from this comment so keep it nice and descriptive.
Remember to remove sections that you don't need or use.

If it does not make sense to have a summary, you can take that out as well.
You would typically keep the summary only if there is a general theme change that can be summarized. If you find
yourself saying the same thing as any of the bullet points in a slightly different way; then this section is not
needed.

Added

• Describe the added features

Changed

• Describe the outwards facing code change

Fixed

• Describe the fix and the bug

Removed

• Describe what was removed and why

Security

• Describe the security issue and the fix

Deprecated

• Describe the part of the code being deprecated and why

Release v10.8.4

Fixed

• Fixing a problem of clients with reducers and/or reactors having problems sometimes reconnecting. This was traced back to how we were managing connected clients. Removing Grain observer mechanism for this notification that seem to hault reconnects for some time.
• Making the AppendedEventsQueues and specific Queue grains live forever, which is the intended behavior as they are there to scale out handling of events.
• Fixing disposal of Queue task in AppendedEventsQueue to only dispose if task is completed and will now wait for its completion on disposal.
• FIxing TaskCompletionSource for AppendedEventsQueue to run continuations asynchronously.

Release v10.8.3

Fixed

• Fixing a problem of clients with reducers and/or reactors having problems sometimes reconnecting. This was traced back to how we were managing connected clients. Removing Grain observer mechanism for this notification that seem to hault reconnects for some time.

Release v10.8.2

Fixed

• Adding the logger factory for the underlying .NET CronicleClient through its options when using ASP.NET Core setup.

Release v10.8.1

Fixed

• Fixing the EventSequence view to show data again by upgrading Application Model as it holds the dependency to Fundamentals for the Workbench, which includes a bugfix for Guid deserialization (https://github.com/Cratis/Fundamentals/releases/tag/v6.1.1).

Release v10.8.0

Summary

Barebones implementation of configuration for observers. This will be fleshed out later

Added

• Configuration for observers for configuring retry policies using asp net configuration system for the Cratis Server
  • SubscriberTimeout - defaults to 5 seconds
  • MaxRetryAttempts - defaults to 10
  • BackoffDelay - defaults to 1 second
  • ExponentialBackoffDelayFactor - defaults to 2
  • MaximumBackoffDelay - defaults to 600 seconds = 10 minutes