protoactor-dotnet
protoactor-dotnet copied to clipboard
asynkron
Proto Actor - Ultra fast distributed actors for Go, C# and Java/Kotlin
In the remote reader, when we connect to an address and the remote address returns a member-id that is blocked. We should backoff and not try to reconnect within x...
The PartitionActivatorLookup already does this due to the algorithm itself. However, the other ID lookups do not. e.g. PartitionIdentityLookup. let actors continue to stay where they were spawned. We do...
PoC to use spans in serialization API instead of ByteString. I don't expect any real impact, but would be interesting to benchmark and see how it turns out
"GossipResponse System.InvalidOperationException: Did not write as much data as expected." IIRC , this can be the results of writing when someone else is modifying the message
That's true. At present, I have a game server, a room service and a gateway server. The gateway server starts with startclientasync and the room service starts with startmemberasync. Then...
When a system is shutting down, there are a lot of logs that could just be discarded. e.g gossip timeouts, various connection retries etc. which could just be ignored if...
There are cases where it makes sense to skip sending some message to a grain if you know the grain is not active. It would be nice to have some...
Currently, we store an entire ClusterTopology message in the gossip state. This includes Left and Joined members. which are irrelevant here. the only relevant part is topology hash, members and...
Draft for our first community standup ## Date: [To be determined] ## Roadmap Showcase new features * Seed nodes * PartitionIdentityLookup * PartitionActivatorLookup * Cluster Dashboard * OpenTelemetry metrics and...