Async/await first CQRS+ES and DDD framework for .NET
OTHER License
$ dotnet add package EventFlow
EventFlow is a basic CQRS+ES framework designed to be easy to use.
Have a look at our getting started guide, the doβs and donβts and the FAQ.
Development of version 1.0 has started and is mainly braking changes regarding changes
related to replacing EventFlow types with that of Microsoft extension abstractions,
mainly IServiceProvider
and ILogger<>
.
The following list key characteristics of each version as well as its related branches (not properly configured yet).
1.x
Represents the next iteration of EventFlow that aligns EventFlow with the standard packages for .NET (Core). Releases here will only support .NET Standard, .NET Core and .NET versions going forward.
Read the migration guide to view the full list of breaking changes as well as recommendations on how to migrate.
Version 1.x documentation has been pulled into this repository in order to have the code and documentation closer together and (hopefully) have the documentation updated in the same pull-requests as any code changes.
Projects
EventFlow
EventFlow.AspNetCore
EventFlow.Autofac
EventFlow.DependencyInjection
EventFlow.Elasticsearch
EventFlow.EntityFramework
EventFlow.EventStores.EventStore
EventFlow.Hangfire
EventFlow.MongoDB
EventFlow.MsSql
EventFlow.Owin
EventFlow.PostgreSql
EventFlow.Redis
EventFlow.RabbitMQ
EventFlow.Sql
EventFlow.SQLite
EventFlow.TestHelpers
develop-v1
: Development branch, pull requests should be done hererelease-v1
: Release branch, merge commits are done to this branch fromdevelop-v1
to create releases. Typically each commit represents a release0.x
(legacy)
The current stable version of EventFlow and has been the version of EventFlow for almost six years. 0.x versions have .NET Framework support and limited support to the Microsoft extension packages through extra NuGet packages.
Feature and bug fix releases will still be done while there's interest in the community.
develop-v0
: Development branch, pull requests should be done hererelease-v0
: Release branch, merge commits are done to this branch fromdevelop-v0
to create releases. Typically each commit represents a releaseVersion 0.x documentation is (although a bit outdated) is live at https://docs.geteventflow.net/.
List of examples create by different community members. Note that many of these examples will be using EventFlow 0.x.
Create a pull request to get your exampled linked from here.
Racetimes: Shows some features of EventFlow that are not covered in the complete example. It features entities, a read model for an entity, delete on read models, specifications and snapshots.
Racetimes for Azure Functions: Extends the above example to support the HTTP access via Azure Functions
Racetimes for Azure Functions and Event Grid: Further extends the Azure Functions Example to publish to Event Grid, following the RabbitMQ pattern
.NET Core: A Web API running .NET Core 2.2 using the event flow. It uses the pre-defined command/entities/events from the complete example. There are endpoints to create a new example event, getting a data model and to replay all data models.
ElasticSearch/.NET Core: It is configured with EventFlow, ElasticSearch, EventStore, and RabbitMq. See "withRabbitMq" branch for #384.
Vehicle Tracking: A Microservice on .NET Core 2.2 with docker based, you can up the service with docker-compose, this project using various tools to up the services aka. Linux Docker based on .NET Core, RabbitMq, EntityFramework with SQL Server and using EventFlow following CQRS-ES architecture and all microservice can access through ApiGateway which using Ocelot
RestAirline: A classic DDD with CQRS-ES, Hypermedia API project based on EventFlow. It's targeted to ASP.NET Core 2.2 and can be deployed to docker and k8s.
Full Example: A console application on .NET Core 2.2. You can up the services using docker-compose file. Docker-compose file include EventStore, RabbitMq, MongoDb, and PostgreSQL. It include following EventFlow concepts:
Here is a list of the EventFlow concepts. Use the links to navigate to the documentation.
Here's a complete example on how to use the default in-memory event store along with an in-memory read model.
The example consists of the following classes, each shown below
ExampleAggregate
: The aggregate rootExampleId
: Value object representing the identity of the aggregate rootExampleEvent
: Event emitted by the aggregate rootExampleCommand
: Value object defining a command that can be published to theExampleCommandHandler
: Command handler which EventFlow resolves using its IoCExampleReadModel
: In-memory read model providing easy access to the currentNote: This example is part of the EventFlow test suite, so checkout the code and give it a go.
[Test]
public async Task Example()
{
// We wire up EventFlow with all of our classes. Instead of adding events,
// commands, etc. explicitly, we could have used the the simpler
// AddDefaults(Assembly) instead.
var serviceCollection = new ServiceCollection()
.AddLogging()
.AddEventFlow(o => o
.AddEvents(typeof(ExampleEvent))
.AddCommands(typeof(ExampleCommand))
.AddCommandHandlers(typeof(ExampleCommandHandler))
.UseInMemoryReadStoreFor<ExampleReadModel>());
using (var serviceProvider = serviceCollection.BuildServiceProvider())
{
// Create a new identity for our aggregate root
var exampleId = ExampleId.New;
// Resolve the command bus and use it to publish a command
var commandBus = serviceProvider.GetRequiredService<ICommandBus>();
await commandBus.PublishAsync(
new ExampleCommand(exampleId, 42), CancellationToken.None);
// Resolve the query handler and use the built-in query for fetching
// read models by identity to get our read model representing the
// state of our aggregate root
var queryProcessor = serviceProvider.GetRequiredService<IQueryProcessor>();
var exampleReadModel = await queryProcessor.ProcessAsync(
new ReadModelByIdQuery<ExampleReadModel>(exampleId), CancellationToken.None);
// Verify that the read model has the expected magic number
exampleReadModel.MagicNumber.Should().Be(42);
}
}
// The aggregate root
public class ExampleAggregate : AggregateRoot<ExampleAggregate, ExampleId>,
IEmit<ExampleEvent>
{
private int? _magicNumber;
public ExampleAggregate(ExampleId id) : base(id) { }
// Method invoked by our command
public void SetMagicNumber(int magicNumber)
{
if (_magicNumber.HasValue)
throw DomainError.With("Magic number already set");
Emit(new ExampleEvent(magicNumber));
}
// We apply the event as part of the event sourcing system. EventFlow
// provides several different methods for doing this, e.g. state objects,
// the Apply method is merely the simplest
public void Apply(ExampleEvent aggregateEvent)
{
_magicNumber = aggregateEvent.MagicNumber;
}
}
// Represents the aggregate identity (ID)
public class ExampleId : Identity<ExampleId>
{
public ExampleId(string value) : base(value) { }
}
// A basic event containing some information
public class ExampleEvent : AggregateEvent<ExampleAggregate, ExampleId>
{
public ExampleEvent(int magicNumber)
{
MagicNumber = magicNumber;
}
public int MagicNumber { get; }
}
// Command for update magic number
public class ExampleCommand : Command<ExampleAggregate, ExampleId>
{
public ExampleCommand(
ExampleId aggregateId,
int magicNumber)
: base(aggregateId)
{
MagicNumber = magicNumber;
}
public int MagicNumber { get; }
}
// Command handler for our command
public class ExampleCommandHandler
: CommandHandler<ExampleAggregate, ExampleId, ExampleCommand>
{
public override Task ExecuteAsync(
ExampleAggregate aggregate,
ExampleCommand command,
CancellationToken cancellationToken)
{
aggregate.SetMagicNumber(command.MagicNumber);
return Task.CompletedTask;;
}
}
// Read model for our aggregate
public class ExampleReadModel : IReadModel,
IAmReadModelFor<ExampleAggregate, ExampleId, ExampleEvent>
{
public int MagicNumber { get; private set; }
public Task ApplyAsync(
IReadModelContext context,
IDomainEvent<ExampleAggregate, ExampleId, ExampleEvent> domainEvent,
CancellationToken _cancellationToken
{
MagicNumber = domainEvent.AggregateEvent.MagicNumber;
return Task.CompletedTask;
}
}
EventFlow is still under development, especially the parts regarding how read models are re-populated.
EventFlow is currently used in production environments and performs very well, but it needs to mature before key APIs are stable.
EventFlow is greatly opinionated, but it's possible to create new implementations for almost every part of EventFlow by registering a different implementation of an interface.
Many of the technical design decisions in EventFlow is based on articles. This section lists some of them. If you have a link with a relevant article, please share it by creating an issue with the link.
EventFlow has several tests that verify that its ability to use the systems it integrates with correctly.
To setup a local test environment run the following commands in the checkout directory of EventFlow.
docker-compose pull
docker-compose up
Alternatively, you can skip the NUnit tests marked with the integration
category.
EventFlow was originally developed in my spare time while I worked at both eBay (2015 to 2021) and Schibsted (2021 and onward).
The MIT License (MIT)
Copyright (c) 2015-2024 Rasmus Mikkelsen
https://github.com/eventflow/EventFlow
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.