Background Tasks in .NET

Background Tasks in .NETIt seems like every application I’ve ever written, at some point needs to run a task in the background that is to be scheduled at some point in the future.

My need for tasks as reoccurring batch jobs has decreased significantly over the years because of my applications being more event driven.

Nowadays, I will often use Event Stores Persistent Subscriptions with my event stream as a queue.  However, I’m not event sourcing all parts of the application and not always using Event Store.

Hangfire

I stumbled upon Hangfire a couple years ago when trying to find solution to running background tasks in a .NET console application running as a service with Topshelf.

When I was trying to find a solution, I need to have tasks distributed across multiple worker services.  Hangfire has this problem solved.

An easy way to perform fire-and-forget, delayed and recurring tasks inside ASP.NET applications. No Windows Service required.

Backed by persistent storages. Open and free for commercial use.

Background method calls and their arguments are serialized and may overcome the process boundaries.

You can use Hangfire on different machines to get more processing power with no configuration – synchronization is performed automatically.

Getting Started

The docs are really good over at the Hangfire site, so I don’t want to rewrite what is already over there.  I do however just want to give you enough to show you how simple it is to use and let you decide if its a viable solution for you.

There are two ways in which you can get started.  If you plan on using Hangfire within an ASP.NET Web Application and want to use SQL Server as your storage mechanism, then you can install the boostrapper package which has everything you need.

Install-Package HangFire -Version 1.5.6

If you plan on using in any other project type (Windows Service, Console, OWIN-Compatible Web Application, Azure Worker Role) or different storage mechanism, you can install the Hangfire.Core.

Install-Package HangFire.Core -Version 1.5.6

Configuration

Configuration is the same if you are in an ASP.NET Web Application or any other project type.  What differs is where you put this configuration.

But it is incredibly straight forward.  You simply use the GlobalConfiguration class to configure the entry point.

Server

The server is what processes the background tasks we will define later.  In order to create a server, you simply create a BackgroundJobServer.  This can be called from any project you wish to act as a Hangfire server.  As with configuration, where you place create the BackgroundJobServerwill depend on the project type.

Create Background Task

There are different types of background tasks that we can now create from any other project type.  These calls do not need to be made from the same application as the server.

  • Fire-and-forget
  • Delay
  • Recurring

Any type of Tasks can be called from any client and it will be executed on a Hangfire server.

Fire-and-forget

You simply call the Enqueue with your action you want to invoke on the server.

The Enqueue method does not call the target method immediately, it runs the following steps instead:

  1. Serialize a method information and all its arguments.
  2. Create a new background job based on the serialized information.
  3. Save background job to a persistent storage.
  4. Enqueue background job to its queue.

For my example, I’m going to Enqueue a new task in the same application as my server.

When you run this application, here is the console output.

Hangfire

Delay & Reoccurring

You can also create tasks that are delayed from the time of creation.  This may be useful for possibly invoking a method to send to a user a couple days after they have signed up for a trial of your software.

You can also create reoccurring tasks and supports CRON expressions.

Data Storage

There are fair amount of stable different storage providers such as Redis, SQL Server, SQL Azure, Mongo, PostgreSql.   Be sure to check out each of the data providers github/website to make sure they are still active and production ready as it seems a few may be out of date or experimental.

I highly recommend you check out this project if you are looking for this type of functionality.

Comments

I’d love to hear how you handle running background tasks. Are you using some other tool?  Please share in the comments below or on twitter.


Validating Commands with the Decorator Pattern

Validating CommandsThis post is how I have manged to handle validating commands and queries and belongs to a series about different aspects of applying CQRS

Although these examples are demonstrating usage in commands, they are aslo applicable to queries.

Examples are using using MediatR to handle our requests.

If you are unfamiliar with CQRS or MediatR, here are some other relevant posts:

Feature

First let’s take a look at our feature for Changing a Customers Pricing Level.  In this example I’ve organized all aspects of this feature into one file.  This single file contains our NancyFX Module and Route, Command and Command Handler.

Validation

One small piece of validation I would like to add, is validating the CustomerId is not an empty Guid.

When deserialization occurs thru the Bind() method, if the CustomerId is not present in the payload, the CustomerId will be an empty Guid (00000000-0000-0000-0000-000000000000).

The simplest approach is to add the validation to our Handler.

This doesn’t look incredibly horrible, however we are violating SRP and as you develop more complex commands, validating them will start to consume your Handler.

Decorator Pattern

In object-oriented programming, the decorator pattern (also known as Wrapper, an alternative naming shared with the Adapter pattern) is a design pattern that allows behavior to be added to an individual object, either statically or dynamically, without affecting the behavior of other objects from the same class.[1] The decorator pattern is often useful for adhering to the Single Responsibility Principle, as it allows functionality to be divided between classes with unique areas of concern.[2]

The decorator pattern is pretty straight forward, and even if you never knew the term, you likely already used it.

Here is our decorator for wrapping the command handler with our new validation handler.

Validation Handler

Now we can extract our validation logic out of our command handler and add it to a validation handler which will be invoked prior to the command handler.

Invoking

The one piece of the puzzle that I can’t show exactly in this example will be dependent on what container you use.  When you create a new instance of MediatR you provide a single and multi instance factory in the constructor.  Generally this will be your preferred container.   Check out our container docs to see how you register a decorator.

Comments

I’d love to hear how you handle command or query validation.  Please share in the comments below or on twitter.


Related Posts

Organize by Feature

Organize by FeatureThis post is about how I organize by feature.  Features being my commands and queries.

If you have not familiar with CQRS, it may be worth a look at some of my other posts to get some Context.

 Layers

What I found interesting was as I moved from thinking about layers to thinking about features and vertical slices, I didn’t immediately organize my code that way.  I would still organize my code by what it represented (eg layer).

My project structure would look something like this:

By Layer

In the example above, for a Web Api type endpoint for returning data regarding a customer, there are 3 files I would have open:

  • Controllers/CustomerController.cs
  • Queries/GetCustomer.cs
  • ViewModels/CustomerViewModel.cs

For commands, the list is slightly reduced to:

  • Controllers/CustomersController.cs
  • Commands/ChangeCustomerPricingLevelc.s

Why?

I was so used to organizing code by layer, but there was no real reason to.  Simply out of habit.

One reason was also the ASP.NET Routing convention.  I always used routes by naming convention rather than by using the Route attribute. I’m not a particular fan of Attributes, but that’s another discussion.

Luckily in NancyFX you define your routes a bit different in the Constructor of your Module (similar to Controller).

Organize by Feature

There is nothing stopping us from taking all relevant classes associated to a command or query and adding them to all one file.  This provides a really easy way to view all the relevant code where you would need to make change.

By organizing by feature, our structure could now look like:

byfeature

Single File

For our ChangeCustomerPricing level command, we can now have the our Module (Controller), Command, Handler all within the same file for easy access and modification.

My convention has been to add them all to the same namspace.  The namespace being the name of the feature.

Our CustomerModel.cs and all other models that may cut across multiple queries or commands will still be organized by Folder.

If we end up adding additional functionality to a command or query, say Validation or Authorization, we can do so in the appropriate feature file.

Tests

One complaint I still have is not being able to have Tests in a very accessible way since they are in a separate test project.  Resharpers “Show Covering Tests”is an easier way to navigate but I’m still looking for something more.

Comments

I’d love to hear how you organize by feature or any other insights.  Please share in the comments below or on twitter.


Related Posts