EasyNetQ - Custom error queue name based on the original queue

Question

I'd like to give error queues names in EasyNetQ based on the name of the queue that originated the error. For example, all faulted messages from QueueA would go to QueueA.Error; QueueB messages would be redirected to QueueB.Error, and so on.

However, the ErrorQueueNamingConvention doesn't receive a MessageReceivedInfo parameter, as ErrorExchangeNamingConvention does, and so I cannot know what is the name of the original queue. Is there any way of getting it or any workarounds?

Thanks

score 2 · Answer 1 · edited May 23 '17 at 12:29

You can do this by setting the ErrorQueueNamingConvention on the IBus. So it's done at the bus level, not the individual message level. I name my error queues to match the queue to which the consumer is bound (which I define in configuration). When errors occur while processing messages from queue A, for example, those errors will be routed to queue A_Errors. Here's a sample:

                var errorExchangeName = _configuration.ExchangeName + "_Errors";
                var errorQueueName = _configuration.ListenerQueueName + "_Errors";

                var conventions = _bus.Advanced.Container.Resolve<IConventions>();
                conventions.ErrorExchangeNamingConvention = info => errorExchangeName;
                conventions.ErrorQueueNamingConvention = () => errorQueueName;

You can see in the following code fragment from https://github.com/EasyNetQ/EasyNetQ/blob/master/Source/EasyNetQ/Consumer/DefaultConsumerErrorStrategy.cs how EasyNetQ uses your ErrorQueueNamingConvention to bind to the error queue so it can forward messages to it:

private string DeclareErrorExchangeAndBindToDefaultErrorQueue(IModel model, ConsumerExecutionContext context)
        {
            var originalRoutingKey = context.Info.RoutingKey;

            return errorExchanges.GetOrAdd(originalRoutingKey, _ =>
            {
                var exchangeName = conventions.ErrorExchangeNamingConvention(context.Info);
                model.ExchangeDeclare(exchangeName, ExchangeType.Direct, durable: true);
                model.QueueBind(conventions.ErrorQueueNamingConvention(), exchangeName, originalRoutingKey);
                return exchangeName;
            });
        }

So, if you need even more control then you can override this DefaultConsumerErrorStrategy as well. There you can get at the context you are looking for, though I don't think you need it to get the behavior you are looking for.

Also see this answer.

EasyNetQ - Custom error queue name based on the original queue

1 Answers1