Host external executables in Aspire
이 콘텐츠는 아직 번역되지 않았습니다.
In Aspire, you can host external executable applications alongside your projects using the AddExecutable method. This capability is useful when you need to integrate executable applications or tools into your distributed application, such as Node.js applications, Python scripts, or specialized CLI tools.
When to use executable resources
Section titled “When to use executable resources”Use executable resources when you need to:
- Host non-.NET applications that don’t have containerized equivalents.
- Integrate command-line tools or utilities into your application.
- Run external processes that other resources depend on.
- Develop with tools that provide local development servers.
Common examples include:
- Frontend development servers: Tools like Vercel CLI or webpack dev server.
- Language-specific applications: Node.js apps, Python scripts, or Go applications.
- Database tools: Migration utilities or database seeders.
- Build tools: Asset processors or code generators.
Basic usage
Section titled “Basic usage”The AddExecutable method requires a resource name, the executable path, and optionally command-line arguments and a working directory:
var builder = DistributedApplication.CreateBuilder(args);
// Basic executable without argumentsvar nodeApp = builder.AddExecutable("frontend", "node", ".", "server.js");
// Executable with command-line argumentsvar pythonApp = builder.AddExecutable( "api", "python", ".", "-m", "uvicorn", "main:app", "--reload", "--host", "0.0.0.0", "--port", "8000");
builder.Build().Run();This code demonstrates setting up a basic executable resource. The first example runs a Node.js server script, while the second starts a Python application using Uvicorn with specific configuration options passed as arguments directly to the AddExecutable method.
Resource dependencies and environment configuration
Section titled “Resource dependencies and environment configuration”You can provide command-line arguments directly in the AddExecutable call and configure environment variables for resource dependencies. Executable resources can reference other resources and access their connection information.
Arguments in the AddExecutable call
Section titled “Arguments in the AddExecutable call”var builder = DistributedApplication.CreateBuilder(args);
// Arguments provided directly in AddExecutablevar app = builder.AddExecutable( "vercel-dev", "vercel", ".", "dev", "--listen", "3000");Resource dependencies with environment variables
Section titled “Resource dependencies with environment variables”For arguments that depend on other resources, use environment variables:
var builder = DistributedApplication.CreateBuilder(args);
var redis = builder.AddRedis("cache");var postgres = builder.AddPostgres("postgres").AddDatabase("appdb");
var app = builder.AddExecutable("worker", "python", ".", "worker.py") .WithReference(redis) // Provides ConnectionStrings__cache .WithReference(postgres); // Provides ConnectionStrings__appdbWhen one resource depends on another, WithReference passes along environment variables containing the dependent resource’s connection details. For example, the worker executable’s reference to redis and postgres provides it with the ConnectionStrings__cache and ConnectionStrings__appdb environment variables, which contain connection strings to these resources.
Access specific endpoint information
Section titled “Access specific endpoint information”For more control over how connection information is passed to your executable:
var builder = DistributedApplication.CreateBuilder(args);
var redis = builder.AddRedis("cache");
var app = builder.AddExecutable("app", "node", ".", "app.js") .WithReference(redis) .WithEnvironment(context => { // Provide individual connection details context.EnvironmentVariables["REDIS_HOST"] = redis.Resource.PrimaryEndpoint.Property(EndpointProperty.Host); context.EnvironmentVariables["REDIS_PORT"] = redis.Resource.PrimaryEndpoint.Property(EndpointProperty.Port); });Practical example: Vercel CLI
Section titled “Practical example: Vercel CLI”Here’s a complete example using the Vercel CLI to host a frontend application with a backend API:
var builder = DistributedApplication.CreateBuilder(args);
// Backend APIvar api = builder.AddProject<Projects.Api>("api") .WithExternalHttpEndpoints();
// Frontend with Vercel CLIvar frontend = builder.AddExecutable( "vercel-dev", "vercel", ".", "dev", "--listen", "3000") .WithEnvironment("API_URL", api.GetEndpoint("http")) .WithHttpEndpoint(port: 3000, name: "http");
builder.Build().Run();Configure endpoints
Section titled “Configure endpoints”Executable resources can expose HTTP endpoints that other resources can reference:
var builder = DistributedApplication.CreateBuilder(args);
var frontend = builder.AddExecutable( "webpack-dev", "npx", ".", "webpack", "serve", "--port", "8080", "--host", "0.0.0.0") .WithHttpEndpoint(port: 8080, name: "http");
// Another service can reference the frontendvar e2eTests = builder.AddExecutable("playwright", "npx", ".", "playwright", "test") .WithEnvironment("BASE_URL", frontend.GetEndpoint("http"));Environment configuration
Section titled “Environment configuration”Configure environment variables for your executable:
var builder = DistributedApplication.CreateBuilder(args);
var app = builder.AddExecutable( "api", "uvicorn", ".", "main:app", "--reload", "--host", "0.0.0.0") .WithEnvironment("DEBUG", "true") .WithEnvironment("LOG_LEVEL", "info") .WithEnvironment(context => { // Dynamic environment variables context.EnvironmentVariables["START_TIME"] = DateTimeOffset.UtcNow.ToString(); });Publishing with PublishAsDockerfile
Section titled “Publishing with PublishAsDockerfile”For production deployment, executable resources need to be containerized. Use the PublishAsDockerfile method to specify how the executable should be packaged:
var builder = DistributedApplication.CreateBuilder(args);
var app = builder.AddExecutable( "frontend", "npm", ".", "start", "--port", "3000") .PublishAsDockerfile();When you call PublishAsDockerfile(), Aspire generates a Dockerfile during the publish process. You can customize this by providing your own Dockerfile:
Custom Dockerfile for publishing
Section titled “Custom Dockerfile for publishing”Create a Dockerfile in your executable’s working directory:
FROM node:22-alpineWORKDIR /appCOPY package*.json ./RUN npm ci --only=productionCOPY . .EXPOSE 3000CMD ["npm", "start"]Then reference it in your AppHost:
var builder = DistributedApplication.CreateBuilder(args);
var app = builder.AddExecutable("frontend", "npm", ".", "start") .PublishAsDockerfile([new DockerfileBuildArg("NODE_ENV", "production")]);Best practices
Section titled “Best practices”When working with executable resources:
- Use explicit paths: For better reliability, use full paths to executables when possible.
- Handle dependencies: Use
WithReferenceto establish proper dependency relationships. - Configure explicit start: Use
WithExplicitStart()for executables that shouldn’t start automatically. - Prepare for deployment: Always use
PublishAsDockerfile()for production scenarios. - Environment isolation: Use environment variables rather than command-line arguments for sensitive configuration.
- Resource naming: Use descriptive names that clearly identify the executable’s purpose.