Stratara.ServiceDefaults 4.4.9

• net10.0

  • Destructurama.Attributed (>= 5.3.0)
  • Microsoft.Extensions.Http.Resilience (>= 10.11.0)
  • Microsoft.Extensions.ServiceDiscovery (>= 10.11.0)
  • OpenTelemetry.Exporter.OpenTelemetryProtocol (>= 1.20.1)
  • OpenTelemetry.Extensions.Hosting (>= 1.20.1)
  • OpenTelemetry.Instrumentation.EntityFrameworkCore (>= 1.20.1-beta.1)
  • OpenTelemetry.Instrumentation.Http (>= 1.20.1)
  • OpenTelemetry.Instrumentation.Runtime (>= 1.20.1)
  • RabbitMQ.Client.OpenTelemetry (>= 1.0.0-rc.2)
  • Serilog.Extensions.Hosting (>= 10.0.0)
  • Serilog.Formatting.Compact (>= 3.0.0)
  • Serilog.Settings.Configuration (>= 10.0.1)
  • Serilog.Sinks.Async (>= 2.1.0)
  • Serilog.Sinks.Console (>= 6.1.1)
  • Serilog.Sinks.File (>= 7.0.0)
  • Serilog.Sinks.OpenTelemetry (>= 4.2.0)
  • Stratara.Shared (>= 4.4.9)

A fix for hosts on the Orleans execution model. A host that removed or renamed a projection no longer has that
projection's store readers come back on every start and fail each entry; they retire once, as a saga's readers do. A
retired store reader now logs its retirement once instead of three times, and leaves no reminder error behind. No
API change, no schema change and no migration.

### Fixed

- **A reader of a removed projection retires.** On the Orleans execution model, a host that removed or renamed a
 projection kept the keep-alive reminders of its store readers. On every start of the silo they brought the readers
 back, and each one read the next entry of its partition, failed it because no projection of its name was
 registered, and tried again on the next tick, without end — a warning per partition and entry. Such a reader now
 retires the first time it is brought back, as a saga's reader already did: it unregisters its keep-alive, logs
 `117_136` `UnregisteredProjectionReaderRetired` (Information) once, reads nothing and deactivates. Its checkpoints
 stay, so a projection registered again resumes where it stopped; *Operate the Orleans execution model* says how to
 delete them. In a rolling deployment that adds a projection, a reader that lands on a silo of the earlier version
 refuses a rebuild's or a replay's pause, which then fails naming it, and a wait counts its partition as behind.
- **A retired store reader logs its retirement once.** A reader that retired as it was activated — of a partition
 beyond a lowered partition count, or of a saga the silo no longer registers — deactivated before it answered the
 call that activated it. Orleans forwarded that call to a new activation, which retired again, so the retirement was
 logged three times per reader and the keep-alive tick ended in `Could not deliver reminder tick` at Error. The
 reader now answers the call first and deactivates after it.