Tutorial: Create a multi-container app with Docker Compose

In this tutorial, you learn how to manage more than one container and communicate between them when using Container Tools in Visual Studio. Managing multiple containers requires container orchestration and requires an orchestrator such as Docker Compose or Service Fabric. For these procedures, you use Docker Compose. Docker Compose is great for local debugging and testing in the course of the development cycle.

The completed sample that you create in this tutorial can be found on GitHub at https://github.com/MicrosoftDocs/vs-tutorial-samples in the folder docker/ComposeSample.

Prerequisites

• Docker Desktop
• Visual Studio 2022 with the Web Development, Azure Tools workload, and/or .NET cross-platform development workload installed. This installation includes the .NET 8 development tools.

Create a Web Application project

In Visual Studio, create an ASP.NET Core Web App project, named WebFrontEnd, to create a web application with Razor pages.

Note

In Visual Studio 2022 17.2 and later, you can use Azure Functions for this project instead.

Don't select Enable Docker Support. You add Docker support later in the process.

Create a Web API project

1. Add a project to the same solution and call it WebAPI. Select API as the project type, and clear the checkbox for Configure for HTTPS. In this design, we're only using SSL for communication with the client, not for communication from between containers in the same web application. Only WebFrontEnd needs HTTPS and the code in the examples assumes that you have cleared that checkbox. In general, the .NET developer certificates used by Visual Studio are only supported for external-to-container requests, not for container-to-container requests.

   

2. Add support for Azure Cache for Redis. Add the NuGet package Microsoft.Extensions.Caching.StackExchangeRedis (not StackExchange.Redis). In Program.cs, add the following lines, just before var app = builder.Build():

   builder.Services.AddStackExchangeRedisCache(options =>
      {
         options.Configuration = "redis:6379"; // redis is the container name of the redis service. 6379 is the default port
         options.InstanceName = "SampleInstance";
      });
   

3. Add using directives in Program.cs for Microsoft.Extensions.Caching.Distributed and Microsoft.Extensions.Caching.StackExchangeRedis.

   using Microsoft.Extensions.Caching.Distributed;
   using Microsoft.Extensions.Caching.StackExchangeRedis;
   

4. In the Web API project, delete the existing WeatherForecast.cs and Controllers/WeatherForecastController.cs, and add a file under Controllers, CounterController.cs, with the following contents:

   using Microsoft.AspNetCore.Mvc;
   using Microsoft.Extensions.Caching.Distributed;
   using StackExchange.Redis;
   
   namespace WebApi.Controllers
   {
       [ApiController]
       [Route("[controller]")]
       public class CounterController : ControllerBase
       {
           private readonly ILogger<CounterController> _logger;
           private readonly IDistributedCache _cache;
   
           public CounterController(ILogger<CounterController> logger, IDistributedCache cache)
           {
               _logger = logger;
               _cache = cache;
           }
   
           [HttpGet(Name = "GetCounter")]
           public string Get()
           {
               string key = "Counter";
               string? result = null;
               try
               {
                   var counterStr = _cache.GetString(key);
                   if (int.TryParse(counterStr, out int counter))
                   {
                       counter++;
                   }
                   else
                   {
                       counter = 0;
                   }
                   result = counter.ToString();
                   _cache.SetString(key, result);
               }
               catch(RedisConnectionException)
               {
                   result = "Redis cache is not found.";
               }
               return result;
           }
       }
   }
   

   The service increments a counter every time the page is accessed and stores the counter in the cache.

Add code to call the Web API

1. In the WebFrontEnd project, open the Index.cshtml.cs file, and replace the OnGet method with the following code.

   public async Task OnGet()
   {
      using (var client = new System.Net.Http.HttpClient())
      {
         // Call *mywebapi*, and display its response in the page
         var request = new System.Net.Http.HttpRequestMessage();
         // webapi is the container name
         request.RequestUri = new Uri("http://webapi/Counter");
         var response = await client.SendAsync(request);
         string counter = await response.Content.ReadAsStringAsync();
         ViewData["Message"] = $"Counter value from cache :{counter}";
      }
   }
   

   Note

   In real-world code, you shouldn't dispose HttpClient after every request. For best practices, see Use HttpClientFactory to implement resilient HTTP requests.

2. In the Index.cshtml file, add a line to display ViewData["Message"] so that the file looks like the following code:

   @page
   @model IndexModel
   @{
       ViewData["Title"] = "Home page";
   }
   
   <div class="text-center">
       <h1 class="display-4">Welcome</h1>
       <p>Learn about <a href="/aspnet/core">building Web apps with ASP.NET Core</a>.</p>
       <p>@ViewData["Message"]</p>
   </div>
   

   This code displays the value of the counter returned from the Web API project.

Add Docker Compose support

1. In the WebFrontEnd project, choose Add > Container Orchestrator Support. The Docker Support Options dialog appears.

2. Choose Docker Compose.

3. Choose your Target OS, for example, Linux.

   

   Visual Studio creates a docker-compose.yml file and a .dockerignore file in the docker-compose node in the solution, and that project shows in boldface font, which shows that it's the startup project.

   

   The docker-compose.yml appears as follows:

   version: '3.4'
   
    services:
      webfrontend:
        image: ${DOCKER_REGISTRY-}webfrontend
        build:
          context: .
          dockerfile: WebFrontEnd/Dockerfile
   

   The version specified in the first line is the Docker Compose file version. You normally shouldn't change it, since it's used by the tools to understand how to interpret the file.

   The .dockerignore file contains file types and extensions that you don't want Docker to include in the container. These files are generally associated with the development environment and source control, not part of the app or service you're developing.

   Look at the Container Tools section of the output pane for details of the commands being run. You can see the command-line tool docker-compose is used to configure and create the runtime containers.

4. In the Web API project, again right-click on the project node, and choose Add > Container Orchestrator Support. Choose Docker Compose, and then select the same target OS.

   Note

   In this step, Visual Studio will offer to create a Dockerfile. If you do this on a project that already has Docker support, you are prompted whether you want to overwrite the existing Dockerfile. If you've made changes in your Dockerfile that you want to keep, choose no.

   Visual Studio makes some changes to your docker-compose YML file. Now both services are included.

   version: '3.4'
   
   services:
     webfrontend:
       image: ${DOCKER_REGISTRY-}webfrontend
       build:
         context: .
         dockerfile: WebFrontEnd/Dockerfile
   
     mywebapi:
       image: ${DOCKER_REGISTRY-}mywebapi
       build:
         context: .
         dockerfile: MyWebAPI/Dockerfile
   

5. Add the cache to the docker-compose.yml file:

   redis:
      image: redis
   

   Make sure the indentation is at the same level as the other two services.

6. The first project that you add container orchestration to is set up to be launched when you run or debug. You can configure the launch action in the Project Properties for the Docker Compose project. On the Docker Compose project node, right-click to open the context menu, and then choose Properties, or use Alt+Enter. For example, you can change the page that is loaded by customizing the Service URL property.

   

7. Press F5. Here's what you see when launched:

   

8. You can monitor the containers using the Containers window. If you don't see the window, use the search box, press Ctrl+K, Ctrl+O, or press Ctrl+Q. Under Feature search, search for containers, and choose View > Other Windows > Containers from the list.

9. Expand the Solution Containers node, and choose the node for your Docker Compose project to view combined logs in the Logs tab of this window.

   

   You can also select the node for an individual container to view logs, environment variables, the filesystem, and other details.

Set up launch profiles

1. This solution has an Azure Cache for Redis, but it's not efficient to rebuild the cache container every time you start a debugging session. To avoid that situation, you can set up a couple of launch profiles. Create one profile to start the Azure Cache for Redis. Create a second profile to start the other services. The second profile can use the cache container that's already running. From the menu bar, you can use the dropdown list next to the start button to open a menu with debugging options. Select Manage Docker Compose Launch Settings.

   

   The Manage Docker Compose Launch Settings dialog comes up. With this dialog, you can control which subset of services is launched during a debugging session, which are launched with or without the debugger attached, and the launch service and URL. See Start a subset of Compose services.

   

   Choose New to create a new profile, and name it Start Redis. Then, set the Redis container to Start without debugging, leave the other set to Do not start, and choose Save.

   

   Then create another profile Start My Services that doesn't start Redis, but starts the other two services.

   

   (Optional) Create a third profile Start All to start everything. You can choose Start without debugging for Redis.

2. Choose Start Redis from the dropdown list on the main Visual Studio toolbar. The Redis container builds and starts without debugging. You can use the Containers window to see that it's running. Next, choose Start My Services from the dropdown list and press F5 to launch them. Now you can keep the cache container running throughout many subsequent debug sessions. Every time you use Start My Services, those services use the same cache container.

Congratulations, you're running a Docker Compose application with a custom Docker Compose profile.

Next steps

Look at the options for deploying your containers to Azure.

See also

Docker Compose

Container Tools