⚡ Saga Pattern in .NET Core

📖 What is Saga Pattern?

The Saga Pattern is used in microservices to manage distributed transactions. Instead of a single ACID transaction across services, it breaks the workflow into smaller steps (local transactions). If one step fails, compensating transactions roll back the previous steps.

🛠 Example in .NET Core

Imagine an Order Service that coordinates with Payment Service and Inventory Service.

    // Order Saga Orchestrator (simplified)
    public class OrderSaga
    {
        public async Task ExecuteAsync()
        {
            try
            {
                await ReserveInventory();
                await ProcessPayment();
                await ConfirmOrder();
            }
            catch (Exception)
            {
                await Compensate();
            }
        }

        private Task ReserveInventory() => Task.Run(() => Console.WriteLine("Inventory Reserved"));
        private Task ProcessPayment() => Task.Run(() => Console.WriteLine("Payment Processed"));
        private Task ConfirmOrder() => Task.Run(() => Console.WriteLine("Order Confirmed"));

        // Compensation logic
        private Task Compensate() => Task.Run(() => Console.WriteLine("Rollback: Cancel Payment & Release Inventory"));
    }
    

✅ Advantages

• Ensures data consistency across microservices without 2PC (two-phase commit).
• Improves fault tolerance with compensating transactions.
• Scales well in distributed systems.
• Supports long-running business processes.

⚠️ Disadvantages

• Complex to design and implement.
• Compensating logic can be tricky and error-prone.
• Eventual consistency (not immediate consistency).
• Harder debugging due to asynchronous flows.

🧭 Best Practices

• Use an Orchestration-based Saga (central coordinator) for complex workflows.
• Use Choreography-based Saga (event-driven) for simpler flows.
• Clearly define compensating transactions for each step.
• Implement idempotency to avoid duplicate operations.
• Use message brokers (e.g., RabbitMQ, Kafka, Azure Service Bus) for reliability.

🔒 Precautions

• Ensure compensating transactions do not cause further inconsistencies.
• Monitor and log each saga step for observability.
• Handle retries and timeouts gracefully.
• Secure inter-service communication.

🎯 Summary

The Saga Pattern is a powerful way to handle distributed transactions in microservices. In .NET Core, it can be implemented using orchestration or event-driven choreography. While it adds complexity, it ensures resilience and consistency in large-scale systems.

🚀 Saga Pattern with MassTransit in .NET Core

📖 What is MassTransit Saga?

MassTransit is a popular .NET library for message-based distributed applications. It provides built-in support for the Saga Pattern, which helps manage long-running workflows across multiple services using state machines and message queues.

🛠 Example: Order Processing Saga

Consider an e-commerce flow: Order Service → Payment Service → Shipping Service. If payment fails, the saga compensates by canceling the order.

    using MassTransit;
    using System;

    // Define Saga State
    public class OrderState : SagaStateMachineInstance
    {
        public Guid CorrelationId { get; set; }
        public string CurrentState { get; set; }
        public string OrderId { get; set; }
    }

    // Define Events
    public class OrderSubmitted
    {
        public string OrderId { get; set; }
    }

    public class PaymentCompleted
    {
        public string OrderId { get; set; }
    }

    public class PaymentFailed
    {
        public string OrderId { get; set; }
    }

    // Define State Machine
    public class OrderStateMachine : MassTransitStateMachine<OrderState>
    {
        public State Submitted { get; private set; }
        public State Completed { get; private set; }
        public State Failed { get; private set; }

        public Event<OrderSubmitted> OrderSubmittedEvent { get; private set; }
        public Event<PaymentCompleted> PaymentCompletedEvent { get; private set; }
        public Event<PaymentFailed> PaymentFailedEvent { get; private set; }

        public OrderStateMachine()
        {
            InstanceState(x => x.CurrentState);

            Event(() => OrderSubmittedEvent, x => x.CorrelateById(m => m.Message.OrderId));
            Event(() => PaymentCompletedEvent, x => x.CorrelateById(m => m.Message.OrderId));
            Event(() => PaymentFailedEvent, x => x.CorrelateById(m => m.Message.OrderId));

            Initially(
                When(OrderSubmittedEvent)
                    .Then(context => context.Instance.OrderId = context.Data.OrderId)
                    .TransitionTo(Submitted));

            During(Submitted,
                When(PaymentCompletedEvent)
                    .TransitionTo(Completed),
                When(PaymentFailedEvent)
                    .TransitionTo(Failed)
                    .Then(context => Console.WriteLine("Compensating: Cancel Order")));
        }
    }
    

✅ Advantages

• Handles long-running workflows across services.
• Built-in state persistence and correlation.
• Supports retries, compensation, and fault tolerance.
• Integrates with message brokers (RabbitMQ, Azure Service Bus, Kafka).

⚠️ Disadvantages

• Increases system complexity.
• Requires reliable message broker setup.
• Eventual consistency (not immediate consistency).
• Debugging distributed sagas can be challenging.

🧭 Best Practices

• Use idempotent consumers to avoid duplicate processing.
• Persist saga state in a reliable store (SQL, MongoDB, etc.).
• Define clear compensating actions for each step.
• Monitor saga execution with logging and tracing.

🔒 Precautions

• Ensure correlation IDs are unique and consistent.
• Handle message retries and poison messages gracefully.
• Secure communication between services and broker.
• Test compensation logic thoroughly.

🎯 Summary

MassTransit makes implementing the Saga Pattern in .NET Core much easier by providing state machines, persistence, and message-driven orchestration. It’s ideal for distributed systems where workflows span multiple services and need resilience.

Comments and Discussion