Building a Microservices Architecture in .NET 8 Using Ocelot API Gateway

Published: (June 18, 2026 at 04:52 AM EDT)
5 min read
Source: Dev.to

Source: Dev.to

Introduction

In modern applications, we often split our system into multiple smaller services called Microservices. Each service is responsible for a specific business function. For example: User Service → Handles users Product Service → Handles products Order Service → Handles orders When we have multiple services, clients need a simple way to access them. Instead of calling each service directly, we use an API Gateway. In this tutorial, we’ll build: User Microservice Product Microservice Ocelot API Gateway By the end of this article, you’ll understand how API Gateway routing works and how Ocelot can help manage communication between microservices. Our architecture will look like this: Browser / Postman | v API Gateway (Ocelot) /
/
v v User API Product API

Instead of calling User API or Product API directly, all requests will go through the API Gateway. Open Visual Studio and create: dotnet new webapi -n UserService

dotnet new webapi -n ProductService

dotnet new webapi -n ApiGateway

Your solution structure should look like: MicroserviceDemo | ├── ApiGateway ├── UserService └── ProductService

Open UserService and create a controller. [ApiController] [Route(“user”)] public class UserController : ControllerBase { [HttpGet] public IActionResult GetUser() { return Ok(new { Id = 1, Name = “John Doe” }); } }

Run the application. Test: https://localhost:7044/user

Response: { “id”: 1, “name”: “John Doe” }

Great! Our User Service is working. Now create a ProductController. [ApiController] [Route(“product”)] public class ProductController : ControllerBase { [HttpGet] public IActionResult GetProduct() { return Ok(new { Id = 100, Name = “Laptop” }); } }

Run Product Service. Test: https://localhost:7220/product

Response: { “id”: 100, “name”: “Laptop” }

Now both microservices are working independently. At this point, the client needs to know: https://localhost:7044/user

https://localhost:7220/product

Imagine having 20 microservices. The frontend would need to manage 20 different URLs. This becomes difficult to maintain. An API Gateway solves this problem by providing a single entry point. Example: https://localhost:5000/gateway/user

https://localhost:5000/gateway/product

The gateway automatically forwards requests to the correct service. Open Package Manager Console inside ApiGateway. Install Ocelot: Install-Package Ocelot

or dotnet add package Ocelot

Ocelot is a popular API Gateway library built specifically for ASP.NET Core. Create a new file: ocelot.json

Add: { “Routes”: [ { “DownstreamPathTemplate”: “/user”, “DownstreamScheme”: “https”, “DownstreamHostAndPorts”: [ { “Host”: “localhost”, “Port”: 7044 } ], “UpstreamPathTemplate”: “/gateway/user”, “UpstreamHttpMethod”: [ “GET” ] }, { “DownstreamPathTemplate”: “/product”, “DownstreamScheme”: “https”, “DownstreamHostAndPorts”: [ { “Host”: “localhost”, “Port”: 7220 } ], “UpstreamPathTemplate”: “/gateway/product”, “UpstreamHttpMethod”: [ “GET” ] } ] }

Let’s understand what this means. When someone calls: /gateway/user

Ocelot forwards the request to: /user

inside User Service. Similarly: /gateway/product

is forwarded to Product Service. Open ApiGateway Program.cs. Replace the code with: using Ocelot.DependencyInjection; using Ocelot.Middleware;

var builder = WebApplication.CreateBuilder(args);

builder.Configuration.AddJsonFile( “ocelot.json”, optional: false, reloadOnChange: true);

builder.Services.AddOcelot();

var app = builder.Build();

await app.UseOcelot();

app.Run();

This tells the application: Load Ocelot configuration Enable API Gateway middleware Route requests using ocelot.json Start: UserService ProductService ApiGateway Now call: https://localhost:5000/gateway/user

Expected Response: { “id”: 1, “name”: “John Doe” }

Call: https://localhost:5000/gateway/product

Expected Response: { “id”: 100, “name”: “Laptop” }

Notice something important. The client never talks directly to User Service or Product Service. Everything goes through the API Gateway. Suppose we want to allow only 4 requests per minute. Update User Route: “RateLimitOptions”: { “EnableRateLimiting”: true, “Period”: ”60s”, “Limit”: 4 }

Now: Request 1 → Success Request 2 → Success Request 3 → Success Request 4 → Success Request 5 → Blocked Response: 429 Too Many Requests

This helps protect APIs from abuse. Suppose User Service is getting heavy traffic. We create another instance running on: https://localhost:7045/user

Update Ocelot: “DownstreamHostAndPorts”: [ { “Host”: “localhost”, “Port”: 7044 }, { “Host”: “localhost”, “Port”: 7045 } ], “LoadBalancerOptions”: { “Type”: “RoundRobin” }

How it works: Request 1 → Server 1 Request 2 → Server 2 Request 3 → Server 1 Request 4 → Server 2 Traffic gets distributed automatically. In this guide, we built a simple Microservices Architecture using .NET 8 and Ocelot API Gateway. We started by creating separate User and Product services, then configured Ocelot to act as a central entry point for routing requests to the appropriate service. Along the way, we learned how API Gateways simplify communication between clients and microservices, how route configuration works, and how features like rate limiting and load balancing can improve the reliability and scalability of an application. While this example is intentionally simple, the same concepts are used in real-world enterprise applications. As systems grow, an API Gateway becomes an important component for managing traffic, securing APIs, and providing a consistent entry point for consumers. If you’re new to Microservices, I recommend experimenting further by adding: JWT Authentication Health Checks Swagger Aggregation Docker Containers Service Discovery Distributed Logging These additions will help you move from a basic microservices setup to a production-ready architecture. I hope this article helped you understand how Ocelot works and how it can be used to build scalable .NET applications. If you have any questions or suggestions, feel free to leave a comment or connect with me.

0 views
Back to Blog

Related posts

Read more »

The Model Doesn't Remember. You Do

Introduction Before I dug into how an LLM works, I assumed each chat stored its memory or context in its own. The moment I realized it was just an array with al...