Allow consumers to react once after all processing retries are exhausted while preserving the base acknowledgement policy. Ref: IT-1033
1.8 KiB
HrynCo.RabbitMq
RabbitMQ publisher and consumer base for HrynCo applications.
Contents
RabbitMqSettings— connection settings record (host, port, user, password, virtual host)IRabbitMqPublisher/RabbitMqPublisher— publishes JSON-serialized messages to a named queueRabbitMqConsumerBase<TMessage, TMessageData>— background service base with connection management, manual ACK/NACK, retry, permanent-validation rejection, and backward-compatible context-aware handlingRabbitMqMessageContext— neutral AMQP delivery metadata (MessageId, type, correlation, routing, headers, and redelivery state)IRabbitMqMessage<TMessageData>— message contract interfaceCorrelationContext— correlation ID carrier
Packaging
This package is intended for reuse through NuGet. The test project is excluded from packing.
Consumer extension points
Existing consumers can keep overriding HandleMessageAsync(message, cancellationToken).
Consumers that need AMQP metadata can instead override
HandleMessageAsync(message, context, cancellationToken). The base class retains
ownership of acknowledgements and retries.
Override TryValidateMessage(...) for application-specific permanent validation.
Returning false nacks the delivery without requeue before retry processing begins.
Keep validation errors free of credentials and sensitive payload values.
Override HandleMessageRetriesExhaustedAsync(...) when a consumer must react once to a
terminal processing failure, for example by publishing a neutral failure result to the
requesting client. The hook runs after the final handler exception and before the original
delivery is nacked without requeue. Hook failures are logged and do not replace the
original failure or acknowledgement policy. Application shutdown cancellation does not
invoke the hook.