Aspire 13.5.4 Stops Kafka Health Checks From Leaking a Producer on Every Probe
Before Aspire 13.5.4, every AppHost health check on an AddKafka resource created a new Confluent producer that nobody disposed, so polling threads piled up and the AppHost burned CPU. The fix, and the HealthCheckRegistration.Factory trap behind it, applies to your own health checks too.