Відео

In a requestresponse message where is the replyTo set? in the requestclient? in the consumer? both? thanks in advance

Thanks for your walk-through with MassTransit, I noticed the fork link you uploaded on github don't have changes you made in this video.

Awesome discussion going on. I will be using MassTransit and in place of IMediatR hoping it will work great :)

Amazing! 🙏 Filip is an incredible talent.

Thanks Chris for uploading the video. As you mentioned, this project is the beginning of a UI for MassTransit SQL transport, and we're definitely open for PRs. I'll be putting in more work on it soon. And yeah, dark theme is there, I just need to add switch button in some of future releases.

Can we get such dashboard for other queues like RabbitMQ and Kafka?

Nnnnnnnoice!

Hi Chris, great work. Would it be possible to add an overload to allow the option to set a routing key when adding a recurring job? The publish allows it but the new wrappers don't have a way of passing the values in. Thanks again

Despite the masstransit interfaces are simple or ef/nhibernate have abstractions that implement unit of work, they are still vendor-dependant and I prefer to keep my domain and application layer free of vendor code

Thank you very much! I can't thank you enough!

For all of this to work, is the Saga required? I have a web api that publishes to a Consumer, and then a Handler picks up the message and does the work in another service. This works. However, when the consumer service publishes another message, it does not get consumed by the secondary consumer. I have no clue why this isn't working. Any tips?

Thank you Chris! It is awesome!

Thank you for this functionality. I will give it a try. Is it possible for you to briefly touch on those “issues” with ReplyTo that you talked about in the video?

1) Is there any difference in performance between the old and new approach? Or we just not wasting time creating a temporary queue anymore? 2) I have to turn on new feature both at client and server, right? What if part of the clients will be without this feature?

In case anyone missed, he has a setup which automatically deletes everything on the broker before reconfiguring 1:19

13:45 line 33 await context.Execute(routingSlip) 1、does this two activitys will run in parallel or in order ? 2、if AllocationInventory activity faulted, does PaymentActivity will hit its compensation ?

Not mentioned in this video, but I see Job Distribution Strategies were delivered in 8.3.0. This is huge for folks building multi-tenant apps that unfortunately have to route jobs for some clients to specific machines. I'm really excited to try it. Thanks!

Very nice stuff! 🙂 When is this going to land on main branch? 🙂

Is that the most simple and minimal way to set up mass transit when I don’t want to hook it up to a transport and just use the outbox?

I recently added MassTransit to my toolkit, and I’m really happy to see these videos. Great work, and thank you for everything especially bringing SQL Transports for smaller projects! In addition, how would you requeue failed messages after you fixed the problem in the consumer from SQL Transport?

Can recurring jobs be used with Azure servicebus or only with Postgres?

Amazing library but so complicated to use it.

Please lower your noise cancelling on your recording. It was hard to listen to.

Cool stuff! This is probably capable of scheduling messages in single digit second frequencies, I assume? I'm currently running a selfmade redis round robin lock-based scheduler, but I would like to replace that one. I didn't use Quartz yet, but Hangfire did fail miserably at frequencies of <5 seconds, hence the question.

is sql server supported in the transactinal outbox feature?

Can we have two Initially in our state machine Chris if we can have different scenarios in one saga? Thanks for your awesome tutorials!

Seems like I still did not achieved the level of enlightenment necessary to understand this Sagas stuff 😂Great guru Chris achieved high state of enlightenment I assume.

👍👍🔥🔥

This is very helpful. I know you were showing events generated from the api itself. If you wanted to do something like this to have a out proc cqrs system. Then have a application services consume the data and then fire off the commands to Kafka topic and then Command Handlers to consume and then Event process starts. What are you thoughts on that?

Great video! One question: What would you see in the Aspire tracing when you add the MassTransit tracing to the ServiceDefaults? And also for the MassTransit metrics of course?

We are currently using a git submodule instead. Works just as well for us.

I downloaded the Sample from github and upgraded to .NET 8.0. I find the Queues and Topics being created, the I see the Temp Queues being created, but no Messages are exchanged. If I use a direct Message Retrieval off a Queue using Service Bus Nugets, that works, but the MassTransit code does not work as expected. Your video does demonstrate the message exchange, so I am not clear what the problem is.

Is there any content about the usage of masstransit with topic queues ?

great feature 👍

Chris thank you so much for this content, I don't care if microsoft are trying to add their own abstractions for messaging, MassTransit will always be my go to. You have thought of everything!

Thank you for providing the video content. It was hard to grasp the docs at first but I am 100% on board now after a couple months, and absolutely love MT!

I tried to implement it but kept receiving a "task was canceled" error. Sometimes it works, sometimes I get this error

what if i need to use request response pattern? i use C# but i can also use python or other languages.

Hi Chris, I was thinking how would the persistence layer act when it has a stream of v1 events already stored and now it needs to be able to process v2 events (think field renamed, field deleted, field added, field split, field joined - type of scenarios). In absence of a new explicit type it would be tricky to distinguish between v1 vs v2 event handling and re-hydrate the state of an Aggregate / Business Object. Which means apart from a Contracts package version increment I'd probaby treat the old event types and their schema as immutable, unless the underlying consumers use non append-only storage mechanisms (aka not event sourcing). If you have desire & time to exand these cases on a v2 video, I'd be infinitely thankful.

Very useful, thank you

I've been using MassTransit for a couple of years now. I've always suffixed messages to make sure it's clear what message type they are. Like: *Command *Event *Request *Response. I see you're never doing that, is that intentionally? I'm also always adding a CorrelationId to every message type, is there any good reason to not always include that?

So, if I am understanding this correctly, your 'OrderRoutes' class builds a public http api that in turn uses the RequestClient to 'submit' the order, and then returns an actual Accepted or Rejected response from the Consumer (using ResponseClient) to the api caller. This is an alternative to the usual design where the http api would 'publish' the order to the bus and return a simple 'OK'. Then the api caller is responsible for separately calling the 'get' api to see if the order was actually Accepted or Rejected. Is that a reasonable take-away? It makes sense at the edge (the public http api) but am I correct that this would make less sense once the message is hopping along from consumer to consumer (as in a saga)?

Is it possible for downstream consumers to subscribe to response messages?

Thank you for the clear explanations. it is a little confusing to have a request client send commands when there is a clear distinction between requests and commands so appreciate bringing it up.

Chris Patterson, Do you have any video showing how to deliver same message to multiple consumers using Azure Service Bus?

Hi Chris, any video on major updates of contracts coming?

I'm missing how you should handle the scenario when a consumer receives a message that has an enum value that is not understood. Sure enough, when you upgrade everything like you did, it works, but a common situation is that you deploy a service that emits values that might not be compatible to an older contract (with regards to the enum values). There are ways to get around it, I know, but it would be interesting to hear your take on it. Thanks a lot for your videos!

Thanks for the good content as always! Seconding the interest in videos about feature management

8:38 . The condition is "if(!string.IsNullOrEmpty(context.Message.PurchaseOrderNumber) - thats why u did not get the warning message