Service A -
Have a Rest API to handle the Command-A , on the controller I do some basic field validation (null check , email format check etc).
After validation controller send the Command-A to Command-Handler-A.
Service B - also have the same way of Command Handling like service A.
Now , Service B have a saga which on some event send the Command-A to Service A.
Service A handles that Command-A using the Command-Handler-A.
But in this case Command-A by pass the basic validation which I put on controller of Service-A.
Question -
Is I have to put all the validation in Command Handler/Domain Model ? Or
I have to call the Rest API of Service-A from Service-B saga. (Which make a single entry point in service). Or
Any other solution ?
Using - Axon Framework 4.2 and Axon Server with Spring Boot.
Where to put basic validation on controller or in domain model
Service-A is responsible for ensuring its own pre-conditions.
One of those pre-conditions is that it received a correct message. In the case of Controller-A, you should expect to have a parse that converts the bytes that were received into a validated Command-A. That function might be implemented by the controller itself, or it might be delegated to some other component -- but it will certainly be within the application component, not in the domain model.
Related
I have developed an application which was MVC application. It has a requirement that the application will return json data for one get request.
So I have added apicontroller and created a get method to return json data.
So far so good. but then I thought, is it really needed to add apicontroller to create just one get method.
I started exploring and googling what is the difference other than content negotiation. Got lots of answers and articles but non of them were satisfactory.
So here is the actual confusion, why can't I just create a method in the MVC controller with JsonResponse and return the json data(Which I know only is need for my requirement, but other application on different domain will consume it).
Can anyone convince me why should I use apicontroller instead of MVC JsonResponse for my requirement or should I not be using apicontroller at all.
apology if there is any mistake.
If I get it right the question is Can we use MVC action to serve json content answer is yes! Is it okay to use Json Result? answer is It depends where do you want to consume it
Say I am an in a Web Environment where I have no need for the APIs (that means I am not going to serve my data to multiple clients) If that's the scenario where only your View is going to consume data returned from your Action Method you are good to go. An Action returning a Json Result is basically an Action Result and that's what it is made for.
but If you are in a REST scenario and you need your backend to serve your data to the client de facto standard is to use an independent Web API for that.
Controllers' main responsibility should be to work as an intermediary between your View and Model and whatever service layer you want to bring inside it. on the other hand, Web APIs are data-driven there only purpose is to serve data (use them if you need them)
Web APIs are good cause they give you the flexibility of serving the data to possibly any client that might need it. That's what I would pick if I am starting from scratch but if I only need to serve data to one client Controller Action methods will be way to go.
Hope this helps.
I have a group of microservices, called "client-Foo", where Foo is the name for some particular third-party client.
Using those names as-is creates some really ugly endpoint urls, so I want to translate them to a much nicer hierarchal form.
I added a custom PatternServiceRouteMapper that take the serviceId client-Foo and turns it into the url client/Foo.
This gives the url I want, but also breaks the service mapping because it also changes the registered serviceId to client/Foo; thus, when Zuul goes to route it fails because there is no client/Foo service, its id is client-Foo!
I cannot hardcode any paths because the application requires an arbitrary number of different "client-*" services.
By looking at ZuulProxyAutoConfiguration, you can see that you can replace some beans to achieve your goal and in particular you should consider providing your own implementation of RouteLocator interface or extend DiscoveryClientRouteLocator class.
Your service instances could also register in Eureka server with additional data in metadataMap that your RouteLocator could use.
I would simply add a configuration defined Zuul route :
zuul:
routes:
client-foo: /foo/**
Also I would advise against having a dash in a service-id as it can confuse the config server api (in /config/foo-profile.yml where profile is the spring profile for which you want to get the config).
Lets say I have the Hello World web service defined, from the Service Stack examples, but instead of just calling it as /hello/{name}, I want to be able to call it as /hello/{name}/id, where I define id to be a separate parameter (by storing it in refId, within a CustomUserSession).
In other words, I want to split the authentication and the actual web service call into 2 parts, but I don't want to expose this id in the actual web service call, because this is private, only within my system. How would I go about doing this? My ideal workflow is:
client calls web service : /hello/{name}
authentication happens. As part of this authentication, I get hold of this secret id
I now call the web service as /hello/{name}/id
It's almost as if I want to call a web service from within another web service.
I apologize if this seems like a misguided question, but how can I go about achieving this workflow?
Thanks
I recommend taking a look at ServiceStack routing. Your service will only be called if it matches an existing route (or uses a pre-defined route) which means if you only had:
[Route("/hello/{name}")]
It wont match /hello/{name}/{id} and your service wont be called. To match that route you can either have an explicit route for it, e.g:
[Route("/hello/{name}/{id}")]
But this is public info which you're saying you would like to hide. The alternative is to use a wildcard route, e.g:
[Route("/hello/{name*}")]
Which matches any /hello/* route inc. /hello, /hello/foo, /hello/foo/id, /hello/foo/id/and/more/etc.
But what gets populated in your Hello.Name property would be foo/id which you'd have to detect and parse in your service.
An alternative is to supply the hidden id in a QueryString or POST'ed form parameter which you can use in ServiceStack to further populate any request DTO, e.g. using the original route:
[Route("/hello/{name}")]
Will let you call the service with /route/foo?id=secret which if your Request DTO had an id property will be populated with secret. If you didn't have (or want) to have an id property you can still access the param with:
base.Request.QueryString["id"];
base.Request.FormData["id"];
But since you want this to be hidden, you should probably avoid passing it in the url which is visible by everyone (inc. web proxies, http logs, and other middle ware). A better place to hide a param is to pass it as a custom HTTP header, e.g: X-id: secret which you can access in your service with:
base.Request.Headers["X-id"];
I am creating a WCF data service on top of a EF 4.1 code first data model that goes against a multi-tenant database. In order to enforce the rules of accessing the multi-tenancy I want to require a string id (that is required by all of our MVC-based REST services) be passed as part of the url route. So, for example, right now I have a service route like this:
http://mysample.net/mysamplesvc.svc/Users
That returns all users in the db w/o a filter. I want to partition this by client by requiring the client id be passed as part of the request like this:
http://mysample.net/mysamplesvc.svc/client123/Users
If it is not passed-in or is invalid I will handle it as required. However, I do not see any example like this anywhere. I think I must be searching incorrectly as this seems like a pretty common scenario.
TIA!
The sample service at http://services.odata.org/(S(readwrite))/OData/OData.svc/ does this (it creates a new instance of the data for each (S(...)) in the URL).
The service code is the last sample on this page: http://www.odata.org/ecosystem#samplecode
Is it possible to have a bi-directional web part connection? I am aware that a web part can be both a provider and a consumer but it seems only one connection is allowed between two web parts.
What I am trying to accomplish is a bi-directional connection where a field in web part A can update web Part B and a field in B can update web part A.
You can do this in several ways.
The main things to understand are:
You're in complete control of the interface the web parts share
By default only the consumer knows anything about the other web part
Here is then a couple of ways to implement what you describe:
Option 1: Pull/Push:
You can make your interface such that the consumer can pull the information it needs by either calling function or getting property values
And the consumer also pushes the information the other webpart might need by calling functions or setting property values.
Option 2: Consumer announcing
Here you define your interface (and way of using it) such that when a consumer get's a connection it make a call back through the interface to give the producer a reference to the consumer. This reference can then implement the same or another interface any way out want.
Now both the consumer and the producer has a reference to another web part where they can get the information they need.