Saga Pattern

1️⃣ What Problem Does Saga Solve?

• In a microservices system, one business process (like Place an Order) may need to update multiple services:

  • Order Service creates an order

  • Payment Service charges the customer

  • Inventory Service reserves the item

• You can’t use a single database transaction across all these independent services.

• Challenge: If one step fails, how do you keep data consistent?

✅ Saga Pattern is the solution.
It coordinates a series of local transactions with the ability to undo (compensate) when something fails.

---

2️⃣ Key Idea

• Each microservice does its own local transaction.

• If a later step fails, compensating actions roll back previous steps.

Think of booking a trip:

• Book Flight → Book Hotel → Book Taxi

• If the hotel booking fails, you cancel the flight.

---

3️⃣ Two Main Saga Styles

A. Choreography (Event-Driven)

• No central controller.

• Services communicate through events.

• Each service listens for events and reacts.

Flow:

1. Order Service saves order → publishes OrderCreated.

2. Payment Service listens → charges payment → publishes PaymentDone.

3. Inventory Service listens → reserves stock → publishes InventoryReserved.

4. If any step fails, a compensating event is published (e.g., PaymentFailed → Order Service cancels the order).

✅ Good for simple workflows
❌ Hard to track when flows become complex.

---

B. Orchestration (Central Coordinator)

• A central Saga Orchestrator service tells each service what to do next.

• Orchestrator sends commands and handles success/failure responses.

Flow:

1. Orchestrator → tells Order Service: “Create order.”

2. After success → tells Payment Service: “Charge payment.”

3. After success → tells Inventory Service: “Reserve stock.”

4. If payment fails → tells Order Service: “Cancel order.”

✅ Easier to manage complex processes
❌ Single orchestrator can become a bottleneck if not scaled.

---

4️⃣ Saga in .NET (Common Tools)

Message Broker

• RabbitMQ

• Azure Service Bus

• Kafka

Libraries

• MassTransit: .NET library that supports Saga state machines.

• NServiceBus: Enterprise-grade messaging with saga support.

• Dapr Workflow (for orchestration).

---

5️⃣ Example Flow with MassTransit (Simplified)

1. Define Events/Commands

   • OrderCreated, PaymentCompleted, InventoryReserved.

2. Saga State Machine

   • Tracks the order state and next steps.

3. Services

   • Order API, Payment API, Inventory API listen to commands/events.

4. Compensation

   • If PaymentFailed, send CancelOrder to revert.

---

6️⃣ Benefits

• ✅ Data consistency across microservices without a single database transaction.

• ✅ Flexible—works with event-driven systems.

• ✅ Scalable—each service remains independent.

---

7️⃣ Things to Watch Out For

• More infrastructure: need a message broker and reliable event handling.

• Proper error handling and retries are crucial.

• Ensure idempotency (re-processing an event should not double-charge).

---

🔑 Quick Recap

• Saga = sequence of local transactions + compensating actions.

• Choreography: services react to events (no central brain).

• Orchestration: a coordinator tells services what to do.

• Perfect for .NET microservices where multiple services must stay consistent without using distributed transactions.