Volver al ranking

dotnetcore/CAP

C#cap.dotnetcore.xyz

Distributed transaction solution in micro-service base on eventually consistency, also an eventbus with Outbox pattern

kafkarabbitmqdistributed-transactionsoutbox-patterneventbusmicroservicesservice-bus
Crecimiento de estrellas
Estrellas
7.1k
Forks
1.3k
Crecimiento semanal
Issues
5
2k4k6k
dic 2016feb 2020may 2023jul 2026
README

CAP Logo

CAP                     中文

Docs & Dashboard AppVeyor NuGet NuGet Preview Member project of .NET Core Community GitHub license

CAP is a .NET library that provides a lightweight, easy-to-use, and efficient solution for distributed transactions and event bus integration.

When building SOA or Microservice-based systems, services often need to be integrated via events. However, simply using a message queue cannot guarantee reliability. CAP leverages a local message table, integrated with your current database, to solve exceptions that can occur during distributed system communications. This ensures that event messages are never lost.

You can also use CAP as a standalone EventBus. It offers a simplified approach to event publishing and subscribing without requiring you to inherit or implement any specific interfaces.

Key Features

  • Core Functionality

    • Distributed Transactions: Guarantees data consistency across microservices using a local message table (Outbox Pattern).
    • Event Bus: High-performance, lightweight event bus for decoupled communication.
    • Guaranteed Delivery: Ensures messages are never lost, with automatic retries for failed messages.
  • Advanced Messaging

    • Delayed Messages: Native support for publishing messages with a delay, without relying on message queue features.
    • Flexible Subscriptions: Supports attribute-based, wildcard (*, #), and partial topic subscriptions.
    • Consumer Groups & Fan-Out: Easily implement competing consumer or fan-out patterns for load balancing or broadcasting.
    • Parallel & Serial Processing: Configure consumers for high-throughput parallel processing or ordered sequential execution.
    • Backpressure Mechanism: Automatically manages processing speed to prevent memory overload (OOM) under high load.
  • Extensibility & Integration

    • Pluggable Architecture: Supports a wide range of message queues (RabbitMQ, Kafka, Azure Service Bus, etc.) and databases (SQL Server, PostgreSQL, MongoDB, etc.).
    • Heterogeneous Systems: Provides mechanisms to integrate with non-CAP or legacy systems.
    • Customizable Filters & Serialization: Intercept the processing pipeline with custom filters and support various serialization formats.
  • Monitoring & Observability

    • Real-time Dashboard: A built-in web dashboard to monitor messages, view status, and manually retry.
    • Service Discovery: Integrates with Consul and Kubernetes for node discovery in a distributed environment.
    • OpenTelemetry Support: Built-in instrumentation for distributed tracing, providing end-to-end visibility.

Architecture Overview

CAP Architecture

CAP implements the Outbox Pattern as described in the eShop on .NET ebook.

Getting Started

1. Installation

Install the main CAP package into your project using NuGet.

PM> Install-Package DotNetCore.CAP

Next, install the desired transport and storage providers.

Transports (Message Queues):

PM> Install-Package DotNetCore.CAP.Kafka
PM> Install-Package DotNetCore.CAP.RabbitMQ
PM> Install-Package DotNetCore.CAP.AzureServiceBus
PM> Install-Package DotNetCore.CAP.AmazonSQS
PM> Install-Package DotNetCore.CAP.NATS
PM> Install-Package DotNetCore.CAP.RedisStreams
PM> Install-Package DotNetCore.CAP.Pulsar

Storage (Databases):

The event log table will be integrated into the database you select.

PM> Install-Package DotNetCore.CAP.SqlServer
PM> Install-Package DotNetCore.CAP.MySql
PM> Install-Package DotNetCore.CAP.PostgreSql
PM> Install-Package DotNetCore.CAP.MongoDB     // Requires MongoDB 4.0+ cluster

2. Configuration

Configure CAP in your Startup.cs or Program.cs.

public void ConfigureServices(IServiceCollection services)
{
    // If you are using EF as the ORM
    services.AddDbContext<AppDbContext>(); 
    
    // If you are using MongoDB
    services.AddSingleton<IMongoClient>(new MongoClient("..."));

    services.AddCap(x =>
    {
        // Using Entity Framework
        // CAP can auto-discover the connection string
        x.UseEntityFramework<AppDbContext>();

        // Using ADO.NET
        x.UseSqlServer("Your ConnectionString");
        x.UseMySql("Your ConnectionString");
        x.UsePostgreSql("Your ConnectionString");

        // Using MongoDB (requires a 4.0+ cluster)
        x.UseMongoDB("Your ConnectionString");

        // Choose your message transport
        x.UseRabbitMQ("HostName");
        x.UseKafka("ConnectionString");
        x.UseAzureServiceBus("ConnectionString");
        x.UseAmazonSQS(options => { /* ... */ });
        x.UseNATS("ConnectionString");
        x.UsePulsar("ConnectionString");
        x.UseRedisStreams("ConnectionString");
    });
}

3. Publish Messages

Inject ICapPublisher into your controller or service to publish events. As of version 7.0, you can also publish delayed messages.

public class PublishController : Controller
{
    private readonly ICapPublisher _capBus;

    public PublishController(ICapPublisher capPublisher)
    {
        _capBus = capPublisher;
    }

    [Route("~/adonet/transaction")]
    public IActionResult AdonetWithTransaction()
    {
        using (var connection = new MySqlConnection(ConnectionString))
        {
            // Start a transaction with auto-commit enabled
            using (var transaction = connection.BeginTransaction(_capBus, autoCommit: true))
            {
                // Your business logic...
                _capBus.Publish("xxx.services.show.time", DateTime.Now);
            }
        }
        return Ok();
    }

    [Route("~/ef/transaction")]
    public IActionResult EntityFrameworkWithTransaction([FromServices] AppDbContext dbContext)
    {
        using (var trans = dbContext.Database.BeginTransaction(_capBus, autoCommit: true))
        {
            // Your business logic...
            _capBus.Publish("xxx.services.show.time", DateTime.Now);
        }
        return Ok();
    }

    [Route("~/publish/delay")]
    public async Task<IActionResult> PublishWithDelay()
    {
        // Publish a message with a 30-second delay
        await _capBus.PublishDelayAsync(TimeSpan.FromSeconds(30), "xxx.services.show.time", DateTime.Now);
        return Ok();
    }
}

4. Subscribe to Messages

In a Controller Action

Add the [CapSubscribe] attribute to a controller action to subscribe to a topic.

public class SubscriptionController : Controller
{
    [CapSubscribe("xxx.services.show.time")]
    public void CheckReceivedMessage(DateTime messageTime)
    {
        Console.WriteLine($"Message received: {messageTime}");
    }
}

In a Business Logic Service

If your subscriber is not in a controller, the class must implement the ICapSubscribe interface.

namespace BusinessCode.Service
{
    public interface ISubscriberService
    {
        void CheckReceivedMessage(DateTime datetime);
    }

    public class SubscriberService : ISubscriberService, ICapSubscribe
    {
        [CapSubscribe("xxx.services.show.time")]
        public void CheckReceivedMessage(DateTime datetime)
        {
            // Handle the message
        }
    }
}

Remember to register your service in Startup.cs:

public void ConfigureServices(IServiceCollection services)
{
    services.AddTransient<ISubscriberService, SubscriberService>();

    services.AddCap(x =>
    {
        // ...
    });
}

Asynchronous Subscriptions

For async operations, your subscription method should return a Task and can accept a CancellationToken.

public class AsyncSubscriber : ICapSubscribe
{
    [CapSubscribe("topic.name")]
    public async Task ProcessAsync(Message message, CancellationToken cancellationToken)
    {
        await SomeOperationAsync(message, cancellationToken);
    }
}

Partial Topic Subscriptions

Group topic subscriptions by defining a partial topic on the class level. The final topic will be a combination of the class-level and method-level topics. In this example, the Create method subscribes to customers.create.

[CapSubscribe("customers")]
public class CustomersSubscriberService : ICapSubscribe
{
    [CapSubscribe("create", isPartial: true)]
    public void Create(Customer customer)
    {
        // ...
    }
}

Subscription Groups

Subscription groups are similar to consumer groups in Kafka. They allow you to load-balance message processing across multiple instances of a service.

By default, CAP uses the assembly name as the group name. If multiple subscribers in the same group subscribe to the same topic, only one will receive the message (competing consumers). If they are in different groups, all will receive the message (fan-out).

You can specify a group directly in the attribute:

[CapSubscribe("xxx.services.show.time", Group = "group1")]
public void ShowTime1(DateTime datetime)
{
    // ...
}

[CapSubscribe("xxx.services.show.time", Group = "group2")]
public void ShowTime2(DateTime datetime)
{
    // ...
}

You can also set a default group name in the configuration:

services.AddCap(x =>
{
    x.DefaultGroup = "my-default-group";  
});

Azure Service Bus Emulator Support

The Azure Service Bus Emulator uses separate ports for AMQP messaging (5672) and the HTTP Admin API (5300). Because CAP uses a single connection string for both the ServiceBusClient (AMQP) and the ServiceBusAdministrationClient (HTTP), it cannot target both ports simultaneously.

To work around this, set AutoProvision to false to skip automatic creation of topics, subscriptions, and rules via the Admin API. You must pre-create the required entities in the emulator's configuration instead.

services.AddCap(x =>
{
    x.UseAzureServiceBus(opt =>
    {
        opt.ConnectionString = "Endpoint=sb://localhost;SharedAccessKeyName=RootManageSharedAccessKey;SharedAccessKey=SAS_KEY_VALUE;UseDevelopmentEmulator=true;";
        opt.AutoProvision = false;
    });
});

Note: When AutoProvision is false, topics, subscriptions, and subscription filter rules must already exist before the application starts. This option is also useful when entities are managed externally (e.g., via Infrastructure as Code).

Dashboard

CAP provides a real-time dashboard to view sent and received messages and their status.

PM> Install-Package DotNetCore.CAP.Dashboard

The dashboard is accessible by default at http://localhost:xxx/cap. You can customize the path via options: x.UseDashboard(opt => { opt.PathMatch = "/my-cap"; });.

For distributed environments, the dashboard supports service discovery to view data from all nodes.

Contribute

We welcome contributions! Participating in discussions, reporting issues, and submitting pull requests are all great ways to help. Please read our contributing guidelines (we can create this file if it doesn't exist) to get started.

License

CAP is licensed under the MIT License.

Repositorios relacionados
pathwaycom/pathway

Python ETL framework for stream processing, real-time analytics, LLM pipelines, and RAG.

PythonPyPIOtherbatch-processingkafka
pathway.com
62.6k1.7k
DataTalksClub/data-engineering-zoomcamp

Data Engineering Zoomcamp is a free 9-week course on building production-ready data pipelines. The next cohort starts in January 2026. Join the course here 👇🏼

Jupyter Notebookdata-engineeringkafka
airtable.com/appzbS8Pkg9PL254a/shr6oVXeQvSI5HuWD
43.9k8.6k
apache/kafka

Apache Kafka - A distributed event streaming platform

JavaMavenApache License 2.0scalakafka
kafka.apache.org
33.3k15.4k
yudaocode/SpringBoot-Labs

一个涵盖六个专栏:Spring Boot 2.X、Spring Cloud、Spring Cloud Alibaba、Dubbo、分布式消息队列、分布式事务的仓库。希望胖友小手一抖,右上角来个 Star,感恩 1024

JavaMavenspring-bootspring-cloud
20.1k6.2k
influxdata/telegraf

Agent for collecting, processing, aggregating, and writing metrics, logs, and other arbitrary data.

GoGo ModulesMIT Licensetelegrafmonitoring
influxdata.com/telegraf
17.7k5.8k
heibaiying/BigData-Notes

大数据入门指南 :star:

JavaMavenhadoophdfs
16.9k4.3k
zhisheng17/flink-learning

flink learning blog. http://www.54tianzhisheng.cn/ 含 Flink 入门、概念、原理、实战、性能调优、源码解析等内容。涉及 Flink Connector、Metrics、Library、DataStream API、Table API & SQL 等内容的学习案例,还有 Flink 落地应用的大型项目案例(PVUV、日志存储、百亿数据实时去重、监控告警)分享。欢迎大家支持我的专栏《大数据实时计算引擎 Flink 实战与性能优化》

JavaMavenApache License 2.0flinkkafka
54tianzhisheng.cn/tags/Flink/
15.1k3.9k
aalansehaiyang/technology-talk

【大厂面试专栏】一份Java程序员需要的技术指南,这里有面试题、系统架构、职场锦囊、主流中间件等,让你成为更牛的自己!

javaspring
offercome.cn
14.7k3.8k
sogou/workflow

C++ Parallel Computing and Asynchronous Networking Framework

C++Apache License 2.0taskinghttp
14.4k2.6k
debezium/debezium

Change data capture for a variety of databases. Please log issues at https://github.com/debezium/dbz/issues.

JavaMavenApache License 2.0change-data-capturekafka-connect
debezium.io
12.9k3k
IBM/sarama

Sarama is a Go library for Apache Kafka.

GoGo ModulesMIT Licensekafkakafka-client
12.5k1.9k
redpanda-data/redpanda

Redpanda is a streaming data platform for developers. Kafka API compatible. 10x faster. No ZooKeeper. No JVM!

C++kafkacpp
redpanda.com
12.4k772