ManagedCode.Orleans.SignalR

Trigger On

• configuring an Orleans SignalR backplane or sending messages from grains
• diagnosing connection, group, or user routing across hosts
• upgrading offline delivery, heartbeat persistence, or client invocation handling

Workflow

1. Choose combined or separate silo/endpoint hosts and align packages, clustering, partition options, and named storage.
2. Configure authenticated connection, user, and group targeting.
3. Keep publishing at an explicit domain-event boundary and decide whether bounded offline retention is sufficient.
4. Exercise live delivery, reconnect replay, queue limits, heartbeat cleanup, and interrupted client invocations.
5. Before upgrades, validate prior persisted payloads with the production serializer and choose an explicit compatibility or maintenance-window strategy.

Deliver

• configured backplane and named storage with a stated durability boundary
• a working publisher and client receiver
• evidence for the required live, offline, restart, and failure behavior

Validate

• dotnet restore and dotnet build resolve the aligned package set
• a real connected client receives a grain-originated message on another host
• user/group isolation, reconnect, queue expiry/overflow, and heartbeat cleanup work
• restart tests preserve queued messages only when durable storage is configured
• cancellation and missing-client completion terminate predictably
• tests keep unrelated cases parallel; isolate destructive restart tests by their owned cluster

Related skills

Related agents