azure-eventhub-dotnet

$npx mdskill add microsoft/skills/azure-eventhub-dotnet

Sends and receives events via Azure Event Hubs using .NET

  • Streams high-throughput event data to and from Azure Event Hubs.
  • Deploys `EventHubProducerClient` for sending events, and `EventProcessorClient` with checkpointing for receiving events.
  • Automatically manages partitioning of events across multiple partitions in the Event Hub.
  • Delivers real-time ingestion and processing of streaming event data to applications.

SKILL.md

.github/skills/azure-eventhub-dotnetView on GitHub ↗
---
name: azure-eventhub-dotnet
description: |
  Azure Event Hubs SDK for .NET. Use for high-throughput event streaming: sending events (EventHubProducerClient, EventHubBufferedProducerClient), receiving events (EventProcessorClient with checkpointing), partition management, and real-time data ingestion. Triggers: "Event Hubs", "event streaming", "EventHubProducerClient", "EventProcessorClient", "send events", "receive events", "checkpointing", "partition".
license: MIT
metadata:
  author: Microsoft
  version: "1.0.0"
  package: Azure.Messaging.EventHubs
---

# Azure.Messaging.EventHubs (.NET)

High-throughput event streaming SDK for sending and receiving events via Azure Event Hubs.

## Installation

```bash
# Core package (sending and simple receiving)
dotnet add package Azure.Messaging.EventHubs

# Processor package (production receiving with checkpointing)
dotnet add package Azure.Messaging.EventHubs.Processor

# Authentication
dotnet add package Azure.Identity

# For checkpointing (required by EventProcessorClient)
dotnet add package Azure.Storage.Blobs
```

**Current Versions**: Azure.Messaging.EventHubs v5.12.2, Azure.Messaging.EventHubs.Processor v5.12.2

## Environment Variables

```bash
EVENTHUB_FULLY_QUALIFIED_NAMESPACE=<namespace>.servicebus.windows.net  # Required: Event Hubs fully qualified namespace
EVENTHUB_NAME=<event-hub-name>  # Required: Event Hub name
BLOB_STORAGE_CONNECTION_STRING=<storage-connection-string>  # Alternative to Entra ID auth
BLOB_CONTAINER_NAME=<checkpoint-container>  # Required: checkpoint container name
EVENTHUB_CONNECTION_STRING=Endpoint=sb://<namespace>.servicebus.windows.net/;SharedAccessKeyName=...  # Alternative to Entra ID auth
AZURE_TOKEN_CREDENTIALS=prod  # Required only if DefaultAzureCredential is used in production
```

## Authentication

```csharp
using Azure.Identity;
using Azure.Messaging.EventHubs;
using Azure.Messaging.EventHubs.Producer;

// Local dev: DefaultAzureCredential. Production: set AZURE_TOKEN_CREDENTIALS=prod or AZURE_TOKEN_CREDENTIALS=<specific_credential>
var credential = new DefaultAzureCredential(
    DefaultAzureCredential.DefaultEnvironmentVariableName
);
// Or use a specific credential directly in production:
// See https://learn.microsoft.com/dotnet/api/overview/azure/identity-readme?view=azure-dotnet#credential-classes
// var credential = new ManagedIdentityCredential();

var fullyQualifiedNamespace = Environment.GetEnvironmentVariable("EVENTHUB_FULLY_QUALIFIED_NAMESPACE");
var eventHubName = Environment.GetEnvironmentVariable("EVENTHUB_NAME");

var producer = new EventHubProducerClient(
    fullyQualifiedNamespace,
    eventHubName,
    credential);
```

**Required RBAC Roles**:
- **Sending**: `Azure Event Hubs Data Sender`
- **Receiving**: `Azure Event Hubs Data Receiver`
- **Both**: `Azure Event Hubs Data Owner`

## Client Types

| Client | Purpose | When to Use |
|--------|---------|-------------|
| `EventHubProducerClient` | Send events immediately in batches | Real-time sending, full control over batching |
| `EventHubBufferedProducerClient` | Automatic batching with background sending | High-volume, fire-and-forget scenarios |
| `EventHubConsumerClient` | Simple event reading | Prototyping only, NOT for production |
| `EventProcessorClient` | Production event processing | **Always use this for receiving in production** |

## Core Workflow

### 1. Send Events (Batch)

```csharp
using Azure.Identity;
using Azure.Messaging.EventHubs;
using Azure.Messaging.EventHubs.Producer;

await using var producer = new EventHubProducerClient(
    fullyQualifiedNamespace,
    eventHubName,
    new DefaultAzureCredential());

// Create a batch (respects size limits automatically)
using EventDataBatch batch = await producer.CreateBatchAsync();

// Add events to batch
var events = new[]
{
    new EventData(BinaryData.FromString("{\"id\": 1, \"message\": \"Hello\"}")),
    new EventData(BinaryData.FromString("{\"id\": 2, \"message\": \"World\"}"))
};

foreach (var eventData in events)
{
    if (!batch.TryAdd(eventData))
    {
        // Batch is full - send it and create a new one
        await producer.SendAsync(batch);
        batch = await producer.CreateBatchAsync();
        
        if (!batch.TryAdd(eventData))
        {
            throw new Exception("Event too large for empty batch");
        }
    }
}

// Send remaining events
if (batch.Count > 0)
{
    await producer.SendAsync(batch);
}
```

### 2. Send Events (Buffered - High Volume)

```csharp
using Azure.Messaging.EventHubs.Producer;

var options = new EventHubBufferedProducerClientOptions
{
    MaximumWaitTime = TimeSpan.FromSeconds(1)
};

await using var producer = new EventHubBufferedProducerClient(
    fullyQualifiedNamespace,
    eventHubName,
    new DefaultAzureCredential(),
    options);

// Handle send success/failure
producer.SendEventBatchSucceededAsync += args =>
{
    Console.WriteLine($"Batch sent: {args.EventBatch.Count} events");
    return Task.CompletedTask;
};

producer.SendEventBatchFailedAsync += args =>
{
    Console.WriteLine($"Batch failed: {args.Exception.Message}");
    return Task.CompletedTask;
};

// Enqueue events (sent automatically in background)
for (int i = 0; i < 1000; i++)
{
    await producer.EnqueueEventAsync(new EventData($"Event {i}"));
}

// Flush remaining events before disposing
await producer.FlushAsync();
```

### 3. Receive Events (Production - EventProcessorClient)

```csharp
using Azure.Identity;
using Azure.Messaging.EventHubs;
using Azure.Messaging.EventHubs.Consumer;
using Azure.Messaging.EventHubs.Processor;
using Azure.Storage.Blobs;

// Blob container for checkpointing
var blobClient = new BlobContainerClient(
    Environment.GetEnvironmentVariable("BLOB_STORAGE_CONNECTION_STRING"),
    Environment.GetEnvironmentVariable("BLOB_CONTAINER_NAME"));

await blobClient.CreateIfNotExistsAsync();

// Create processor
var processor = new EventProcessorClient(
    blobClient,
    EventHubConsumerClient.DefaultConsumerGroup,
    fullyQualifiedNamespace,
    eventHubName,
    new DefaultAzureCredential());

// Handle events
processor.ProcessEventAsync += async args =>
{
    Console.WriteLine($"Partition: {args.Partition.PartitionId}");
    Console.WriteLine($"Data: {args.Data.EventBody}");
    
    // Checkpoint after processing (or batch checkpoints)
    await args.UpdateCheckpointAsync();
};

// Handle errors
processor.ProcessErrorAsync += args =>
{
    Console.WriteLine($"Error: {args.Exception.Message}");
    Console.WriteLine($"Partition: {args.PartitionId}");
    return Task.CompletedTask;
};

// Start processing
await processor.StartProcessingAsync();

// Run until cancelled
await Task.Delay(Timeout.Infinite, cancellationToken);

// Stop gracefully
await processor.StopProcessingAsync();
```

### 4. Partition Operations

```csharp
// Get partition IDs
string[] partitionIds = await producer.GetPartitionIdsAsync();

// Send to specific partition (use sparingly)
var options = new SendEventOptions
{
    PartitionId = "0"
};
await producer.SendAsync(events, options);

// Use partition key (recommended for ordering)
var batchOptions = new CreateBatchOptions
{
    PartitionKey = "customer-123"  // Events with same key go to same partition
};
using var batch = await producer.CreateBatchAsync(batchOptions);
```

## EventPosition Options

Control where to start reading:

```csharp
// Start from beginning
EventPosition.Earliest

// Start from end (new events only)
EventPosition.Latest

// Start from specific offset
EventPosition.FromOffset(12345)

// Start from specific sequence number
EventPosition.FromSequenceNumber(100)

// Start from specific time
EventPosition.FromEnqueuedTime(DateTimeOffset.UtcNow.AddHours(-1))
```

## ASP.NET Core Integration

```csharp
// Program.cs
using Azure.Identity;
using Azure.Messaging.EventHubs.Producer;
using Microsoft.Extensions.Azure;

builder.Services.AddAzureClients(clientBuilder =>
{
    clientBuilder.AddEventHubProducerClient(
        builder.Configuration["EventHub:FullyQualifiedNamespace"],
        builder.Configuration["EventHub:Name"]);
    
    clientBuilder.UseCredential(new DefaultAzureCredential());
});

// Inject in controller/service
public class EventService
{
    private readonly EventHubProducerClient _producer;
    
    public EventService(EventHubProducerClient producer)
    {
        _producer = producer;
    }
    
    public async Task SendAsync(string message)
    {
        using var batch = await _producer.CreateBatchAsync();
        batch.TryAdd(new EventData(message));
        await _producer.SendAsync(batch);
    }
}
```

## Best Practices

1. **Use `EventProcessorClient` for receiving** — Never use `EventHubConsumerClient` in production
2. **Checkpoint strategically** — After N events or time interval, not every event
3. **Use partition keys** — For ordering guarantees within a partition
4. **Reuse clients** — Create once, use as singleton (thread-safe)
5. **Use `await using`** — Ensures proper disposal
6. **Handle `ProcessErrorAsync`** — Always register error handler
7. **Batch events** — Use `CreateBatchAsync()` to respect size limits
8. **Use buffered producer** — For high-volume scenarios with automatic batching

## Error Handling

```csharp
using Azure.Messaging.EventHubs;

try
{
    await producer.SendAsync(batch);
}
catch (EventHubsException ex) when (ex.Reason == EventHubsException.FailureReason.ServiceBusy)
{
    // Retry with backoff
    await Task.Delay(TimeSpan.FromSeconds(5));
}
catch (EventHubsException ex) when (ex.IsTransient)
{
    // Transient error - safe to retry
    Console.WriteLine($"Transient error: {ex.Message}");
}
catch (EventHubsException ex)
{
    // Non-transient error
    Console.WriteLine($"Error: {ex.Reason} - {ex.Message}");
}
```

## Checkpointing Strategies

| Strategy | When to Use |
|----------|-------------|
| Every event | Low volume, critical data |
| Every N events | Balanced throughput/reliability |
| Time-based | Consistent checkpoint intervals |
| Batch completion | After processing a logical batch |

```csharp
// Checkpoint every 100 events
private int _eventCount = 0;

processor.ProcessEventAsync += async args =>
{
    // Process event...
    
    _eventCount++;
    if (_eventCount >= 100)
    {
        await args.UpdateCheckpointAsync();
        _eventCount = 0;
    }
};
```

## Related SDKs

| SDK | Purpose | Install |
|-----|---------|---------|
| `Azure.Messaging.EventHubs` | Core sending/receiving | `dotnet add package Azure.Messaging.EventHubs` |
| `Azure.Messaging.EventHubs.Processor` | Production processing | `dotnet add package Azure.Messaging.EventHubs.Processor` |
| `Azure.ResourceManager.EventHubs` | Management plane (create hubs) | `dotnet add package Azure.ResourceManager.EventHubs` |
| `Microsoft.Azure.WebJobs.Extensions.EventHubs` | Azure Functions binding | `dotnet add package Microsoft.Azure.WebJobs.Extensions.EventHubs` |

More from microsoft/skills

SkillDescription
agent-framework-azure-ai-pyBuild Azure AI Foundry agents using the Microsoft Agent Framework Python SDK (agent-framework-azure-ai). Use when creating persistent agents with AzureAIAgentsProvider, using hosted tools (code interpreter, file search, web search), integrating MCP servers, managing conversation threads, or implementing streaming responses. Covers function tools, structured outputs, and multi-tool agents.
airunway-aks-setupSet up AI Runway on AKS — from bare cluster to running model. Covers cluster verification, controller install, GPU assessment, provider setup, and first deployment. WHEN: \"setup AI Runway\", \"onboard AKS cluster\", \"install AI Runway\", \"airunway setup\", \"deploy model to AKS\", \"GPU inference on AKS\", \"KAITO setup on AKS\", \"run LLM on AKS\", \"vLLM on AKS\", \"set up model serving on AKS\", \"AI Runway controller\".
appinsights-instrumentationGuidance for instrumenting webapps with Azure Application Insights. Provides telemetry patterns, SDK setup, and configuration references. WHEN: how to instrument app, App Insights SDK, telemetry patterns, what is App Insights, Application Insights guidance, instrumentation examples, APM best practices.
applicationinsights-web-tsInstrument browser/web apps with the Application Insights JavaScript SDK (@microsoft/applicationinsights-web). Use for Real User Monitoring (RUM) — page views, clicks, AJAX/fetch dependencies, exceptions, custom events, and browser-side GenAI agent traces correlated to backend OpenTelemetry traces. Covers SDK Loader Script and npm setup, framework extensions (React, React Native, Angular), Click Analytics, telemetry initializers, and OTel GenAI semantic conventions for agent/tool/model spans emitted from the browser.
azure-aiUse for Azure AI: Search, Speech, OpenAI, Document Intelligence. Helps with search, vector/hybrid search, speech-to-text, text-to-speech, transcription, OCR. WHEN: AI Search, query search, vector search, hybrid search, semantic search, speech-to-text, text-to-speech, transcribe, OCR, convert text to speech.
azure-ai-agents-persistent-dotnet|
azure-ai-agents-persistent-java|
azure-ai-anomalydetector-javaBuild anomaly detection applications with Azure AI Anomaly Detector SDK for Java. Use when implementing univariate/multivariate anomaly detection, time-series analysis, or AI-powered monitoring.
azure-ai-contentsafety-javaBuild content moderation applications with Azure AI Content Safety SDK for Java. Use when implementing text/image analysis, blocklist management, or harm detection for hate, violence, sexual content, and self-harm.
azure-ai-contentsafety-py|