agentleFS
Postpartum-genushyacinthus29Postpartum-genushyacinthus29/dotnet-skills/skills/dotnet-azure-functions/SKILL.md

dotnet-azure-functions

Skill13 starsChanged 8 months agoMITView on GitHub

Build, review, or migrate Azure Functions in .NET with correct execution model, isolated worker setup, bindings, DI, and Durable Functions patterns.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use typed activity patterns

    line 51
  2. Very specific:

    Inject dependencies via constructor: - Use ILogger<T> and service interfaces for testability

    line 211
  3. Distinctive:

    Keep orchestrator code deterministic: - No I/O, random, DateTime.Now, or Guid.NewGuid() in orchestrators

    line 212

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Azure Functions for .NET

Trigger On

  • working on Azure Functions in .NET
  • migrating from the in-process model to the isolated worker model
  • adding Durable Functions, bindings, or host configuration

Documentation

References

  • Patterns - Isolated worker patterns, Durable Functions patterns, advanced binding patterns
  • Anti-Patterns - Common Azure Functions mistakes and how to avoid them

Workflow

  1. Use isolated worker model for all new work:
    • In-process model reaches end of support on November 10, 2026
    • Runtime v1.x ends support on September 14, 2026
    • Target .NET 8+ for longest support window
  2. Detect current project shape:
    • Target framework and runtime version
    • Worker model (isolated vs in-process)
    • Binding packages and host configuration
  3. Use standard .NET patterns in isolated model:
    • Normal dependency injection
    • Middleware pipeline
    • IOptions<T> for configuration
    • ILogger<T> for logging
  4. For Durable Functions:
    • Validate orchestration determinism constraints
    • Handle replay behavior correctly
    • Use typed activity patterns
  5. Verify both local and deployment behavior.

Isolated Worker Model Setup

Basic Function with DI

// Program.cs
var host = new HostBuilder()
    .ConfigureFunctionsWebApplication()
    .ConfigureServices(services =>
    {
        services.AddApplicationInsightsTelemetryWorkerService();
        services.ConfigureFunctionsApplicationInsights();
        services.AddSingleton<IMyService, MyService>();
    })
    .Build();

host.Run();

HTTP Trigger Function

public class HttpFunctions(ILogger<HttpFunctions> logger, IMyService myService)
{
    [Function("GetItems")]
    public async Task<IActionResult> GetItems(
        [HttpTrigger(AuthorizationLevel.Function, "get", Route = "items")] HttpRequest req)
    {
        logger.LogInformation("Processing GetItems request");
        var items = await myService.GetItemsAsync();
        return new OkObjectResult(items);
    }
}

Queue Trigger with Options

public class QueueFunctions(ILogger<QueueFunctions> logger, IOptions<ProcessingOptions> options)
{
    [Function("ProcessMessage")]
    public async Task ProcessMessage(
        [QueueTrigger("myqueue", Connection = "AzureWebJobsStorage")] string message)
    {
        logger.LogInformation("Processing message: {Message}", message);
        // Process with retry policy from options
    }
}

Middleware Pattern

Custom Middleware

// Program.cs
var host = new HostBuilder()
    .ConfigureFunctionsWebApplication(builder =>
    {
        builder.UseMiddleware<ExceptionHandlingMiddleware>();
        builder.UseMiddleware<CorrelationIdMiddleware>();
    })
    .Build();

// CorrelationIdMiddleware.cs
public class CorrelationIdMiddleware : IFunctionsWorkerMiddleware
{
    public async Task Invoke(FunctionContext context, FunctionExecutionDelegate next)
    {
        var correlationId = context.Features.Get<IHttpRequestFeature>()?.Headers["X-Correlation-Id"]
            ?? Guid.NewGuid().ToString();

        context.Items["CorrelationId"] = correlationId;

        await next(context);
    }
}

Durable Functions Patterns

Function Chaining

[Function(nameof(ChainOrchestrator))]
public static async Task<string> ChainOrchestrator(
    [OrchestrationTrigger] TaskOrchestrationContext context)
{
    var result1 = await context.CallActivityAsync<string>(nameof(Step1), "input");
    var result2 = await context.CallActivityAsync<string>(nameof(Step2), result1);
    var result3 = await context.CallActivityAsync<string>(nameof(Step3), result2);
    return result3;
}

[Function(nameof(Step1))]
public static string Step1([ActivityTrigger] string input) => $"Step1({input})";

[Function(nameof(Step2))]
public static string Step2([ActivityTrigger] string input) => $"Step2({input})";

[Function(nameof(Step3))]
public static string Step3([ActivityTrigger] string input) => $"Step3({input})";

Fan-Out/Fan-In

[Function(nameof(FanOutFanInOrchestrator))]
public static async Task<int[]> FanOutFanInOrchestrator(
    [OrchestrationTrigger] TaskOrchestrationContext context)
{
    var workItems = await context.CallActivityAsync<string[]>(nameof(GetWorkItems), null);

    // Fan out - process all items in parallel
    var tasks = workItems.Select(item =>
        context.CallActivityAsync<int>(nameof(ProcessWorkItem), item));

    // Fan in - wait for all to complete
    var results = await Task.WhenAll(tasks);

    return results;
}

[Function(nameof(ProcessWorkItem))]
public static int ProcessWorkItem([ActivityTrigger] string item)
{
    // Process item and return result
    return item.Length;
}

Human Interaction Pattern

[Function(nameof(ApprovalOrchestrator))]
public static async Task<string> ApprovalOrchestrator(
    [OrchestrationTrigger] TaskOrchestrationContext context)
{
    var request = context.GetInput<ApprovalRequest>();

    // Send notification
    await context.CallActivityAsync(nameof(SendApprovalRequest), request);

    // Wait for external event with timeout
    using var cts = new CancellationTokenSource();
    var approvalTask = context.WaitForExternalEvent<bool>("ApprovalEvent");
    var timeoutTask = context.CreateTimer(context.CurrentUtcDateTime.AddDays(7), cts.Token);

    var winner = await Task.WhenAny(approvalTask, timeoutTask);

    if (winner == approvalTask)
    {
        cts.Cancel();
        return approvalTask.Result ? "Approved" : "Rejected";
    }

    return "Timed out";
}

Best Practices

  1. Use isolated worker model for new development - Full .NET ecosystem access, middleware support, and longer support lifecycle
  2. Inject dependencies via constructor - Use ILogger<T> and service interfaces for testability
  3. Keep orchestrator code deterministic - No I/O, random, DateTime.Now, or Guid.NewGuid() in orchestrators
  4. Handle sensitive data in activities - Fetch secrets from Key Vault in activity functions, never in orchestrators
  5. Use unique task hub names - Prevent accidental sharing when multiple apps use the same storage
  6. Avoid large inputs/outputs - Serialize to blob storage for large payloads to prevent history bloat
  7. Configure concurrency limits - Set appropriate limits in host.json for resource-intensive functions
  8. Keep SDK and extensions updated - Latest versions include performance improvements and bug fixes

Anti-Patterns to Avoid

Anti-PatternWhy It's BadBetter Approach
Mixing in-process and isolated guidanceIncompatible APIs and patternsChoose one model consistently
Non-deterministic orchestrator codeReplay failures, stuck orchestrationsUse context.CurrentUtcDateTime, no I/O
Large orchestrator inputs/outputsHistory bloat, memory issuesStore large data in blob storage
Shared task hub namesMessage conflicts, stuck orchestrationsUse unique names per app
Secrets in orchestrator historySecurity risk, exposed in logsFetch secrets in activity functions
Blocking calls in async functionsThread pool exhaustionUse await throughout
Missing retry policiesTransient failures cause job lossConfigure retry in bindings or code
Ignoring execution model migrationEOL November 2026 for in-processMigrate to isolated worker model

Deployment Considerations

Linux Consumption Plan Limitations

.NET 10+ apps cannot run on Linux Consumption plan.
Use Flex Consumption plan or App Service for .NET 10+.
.NET 9 is the last version supported on Linux Consumption.

host.json Configuration

{
  "version": "2.0",
  "extensions": {
    "durableTask": {
      "hubName": "MyUniqueTaskHub",
      "maxConcurrentActivityFunctions": 10,
      "maxConcurrentOrchestratorFunctions": 5
    }
  },
  "logging": {
    "applicationInsights": {
      "samplingSettings": {
        "isEnabled": true,
        "excludedTypes": "Request"
      }
    }
  }
}

Deliver

  • correct Functions project setup for the isolated worker model
  • clear binding and host configuration
  • middleware for cross-cutting concerns
  • Durable Functions with proper orchestration patterns
  • migration-safe guidance when upgrading execution models

Validate

  • execution model guidance is consistent (isolated only for new work)
  • orchestrator code is deterministic
  • bindings and host settings match the target runtime
  • large payloads are externalized to blob storage
  • retry policies are configured for transient failures
  • local and deployment behavior are both verified

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.