Video

Hi, yes, IMO the cost of spinning up a new instance of the API and Database for every test method in a class is overkill. This is why we spin up one api/db/etc per test class, hence we need that cleanup process which efectively resets the whole setup to a clean/default state after every test method. Edit just to make clear: if the clean-up process works correctly the data generated by a test should not be there when the next test runs, and every test will run against an clean* database.

Hi, thats an empty marker interface I created in part 1 inside the API

@@vasiliioleinic Oh sorry I must have skipped part 1. Thank you!

There is a currently on-going newsletter completely dedicated to automated testing in .NET which covers even more ground :)

Hi it’s excalidraw.

Hi yes, it's excalidraw :)

What sort of asshole comes and complains to someone who is sharing their knowledge for free that their accent is difficult to understand? You clearly don't have the mental capacity and dedication required to learn to speak a 2nd language. Watch the video slower or, preferably, not at all - and instead reflect on how you can become a less substandard human

Hi, sorry for the issue, can you try again with either MREADY20 or MREADY_20.

Hi, its Rider from Jetbrains.

Hi thx, its VSCode with the excalidraw extension.

Please do video 📹 tutor on drawing tools extension thanks 😊

Hi, first of all thank you for the sub, second can you check again? Seems it got taken down sometime ago, so I just re-uploaded it.

This one was using .net7 at the time of recording.

@@vasiliioleinic thanks! Just to avoid confusion I was talking about your paid course...is it .net7?

@@davidpccode The source code for the course is .NET7. There is WIP to update it to .NET8, however the changes (introduced in .NET 8) have a really small impact on the scope of the course itself. (mostly NuGet packages upgrades)

@@vasiliioleinic Perfect .net 7 is perfect... i'm going to buy your course...thanks!!

Hope you will enjoy it, and as always any feedback is welcome.

Can you elaborate on that a bit?

Yes as mentioned in the video in most simple scenarios you can use a native background service just fine. Now things start getting interesting as time progresses, for example if you deploy multiple instances of your app then you will have to ensure that the background service runs only on one of them, and I don't mind reinventing the bicycle when learning stuff, but for production/live projects you are better off with a library that has that functionality built in, in my humble opinion.

Well you are right it's not going to be a bunch of delegates in the startup code. Most likely everyone will create a class for the middleware/filter and implement it there and just wire it up later on. I'm currently using the MinimalAPI's for a couple of projects in live and have no real issue with it so far.

It can be both ways. If you split your solution using logical boundaries(aka modules in this case) any module can listen to events/messages that come from another module.A system might get deployed as a monolith with only logical boundaries or as a set of services with logical and physical boundaries, the choice is yours. As well as the choice of the provider, if you deploy everything as a single unit then you might opt in to use an in-memory provider, if you go the distributed route you will opt in for smth like rabbitmq.

Also there might be cases where you trigger some heavy background job that requires a lot of processing power so you might opt in to add a consumer on a separate ie machine that will do the heavy lifting there and thus offload.

Will fix that in later videos.

They are created automatically with the provided format, that is if you configured them in app-settings like in the video :) Edit : line 35 i the appsettings.json is where you specify the format of how you want your indexes to be called.

​ @vasiliioleinic Thanks for reply, but there is an error related to ssl certificate raised when I run the project locally: The SSL connection could not be established, see inner exception. ---> System.Security.Authentication.AuthenticationException: The remote certificate is invalid because of errors in the certificate chain: UntrustedRoot , I have searched more about it to be configured in C# project windows, but the problem still, any idea?

Sorry never had that issue myself, will have to take a look once im free and come back to you if and when I find something.

Hi, good question, personally I would go with a separate DbContext for the outbox. At least thats how I went with it in the course on ModularMonoliths Im working on right now 😅 (not the CZcams series) so the answer is opinionated in that regard. Edit: nearly forgot about it : also take into account that MassTransit has a default in-memory Outbox by default, with some limitations of course.

​@@vasiliioleinic I've tried to do this, but something doesn't work for me 🫤 I've created separate OutboxDbContext for MassTransit Inbox/Outbox related stuff, but not a single message gets persisted in a database, so something is definitely wrong 🥲 Can you share any details about the ModularMonoliths course? It seems interesting and I feel this is something that could save me 😄

@@drozdgabriel regarding message persistence, try maybe injecting the RegistrationDbContext and calling the SaveChanges method. If it works you can create a small wrapper around the message publisher endpoint and DbContext to not expose it's internal details then. Edit: sorry for long response for some reason CZcams decided not to notify me on this comment...

Thanks for the feedback, noted about the naming improvements, and indeed the “locking” is a thing I missed here. Regarding the first bit there it is going to be part of a follow up series afterwards as it is a huge undertaking for a simple CZcams series in my opinion :)

Yeah sorry about that, saw that only during post edit… the other latest videos on the channel should have the font increased tho…

All is fine with my resolution and all is readable.

1. An modular monolith is a “microservices-ready” setup, and you should be able to easily decouple a module into a separate microservice at any given time, doing that with in-memory bus, well good luck with that. 2. Scalling issues. 3. You might loose messages that are in memory. And last I’d say fancy management dashboard :P but im not sure what in memory provider you are using so that one might come with one also.

@@vasiliioleinic if you use a proper message broker within your MM - it is for sure makes it very ready for splitting into MS in the future. However what are the performance implications? You can easily use the channel based in memory brokers written from scratch - not very hard at all. You don’t get persistence, but if your entire MM is restarted - I’d argue that persistence of messages can be a double edged sword. MassTransit does make it easier to manage the broker, and with their new Db transport - you get even more options to use SQL server or Postgres as a broker

@@vasiliioleinic I disagree with your reply... IMO RabbitMQ, Kafka etc are overkill for modular monoliths and one should prefer using an in-memory-event/massage-bus for inter-module communication with this type of architecture. Only when your app gets popular to the point that you start seeing problems of scale, you deploy some of the modules as their own separate services (and then it's no more a monolith), you switch to RabbitMQ, Kafka etc. by changing implementation of the event/message-bus adapter. You use modular monolith when you don't have scaling issues - if you do then you start switching to microservices, because you maybe want to run more than one instance of the same module etc. Scale issues are the point when you should switch from monolith to distributed. If one cares about losing messages with an in-memory solution then there are patterns like the inbox/outbox for "at least once" delivery and processing of messages.

My use case in enterprise is to use message bus (azure message bus) is if you ever need to provide event messaging to other services (e.g. integration with separate domain’s system)

Yes for this one specific system yes. Having less code to maintain is what I strive for more often than not. I would use my own if I didn’t use mass transit and relied on my own built infrastructure.

@@vasiliioleinic good decision 👍, maybe worth dropping a text comment for ppl who are new to mass transit

There is a video on metrics coming up. As for the other things mentioned will try to squeeze them in somewhere at some point.

I will take that into account for future videos.

@@vasiliioleinic thank you so much. love your videos

Will take into account to increase it even more for next videos

Hi, this is a older video where I was just figuring out everything, the new ones should be way better. Thx for feed back.