[FEATURE] Encina.AzureEventHubs - Azure Event Hubs Transport

[dlrivada]

Summary

Implement Encina.AzureEventHubs transport for high-throughput event streaming.

Motivation

Azure Event Hubs is essential for:

1. Massive Scale: Millions of events per second
2. Azure Native: First-class Azure integration
3. Kafka Compatible: Kafka protocol support
4. Event Streaming: Real-time analytics pipelines

Use Cases

• Telemetry ingestion
• Real-time analytics
• Event streaming pipelines
• IoT data collection

Proposed Solution

public class AzureEventHubsTransport : IMessageTransport
{
    private readonly EventHubProducerClient _producer;
    private readonly EventProcessorClient _processor;
    
    public async Task PublishAsync<T>(T message, string? partitionKey = null, CancellationToken ct = default)
    {
        var eventData = new EventData(JsonSerializer.SerializeToUtf8Bytes(message));
        
        var options = partitionKey != null 
            ? new SendEventOptions { PartitionKey = partitionKey }
            : null;
        
        await _producer.SendAsync(new[] { eventData }, options, ct);
    }
}

Features

• Partitioned publishing (partition key)
• Consumer groups
• Checkpointing with blob storage
• Batch publishing
• Schema Registry integration

Alternatives Considered

1. Use Azure Service Bus instead: Service Bus is message-queue focused while Event Hubs is stream-focused. Event Hubs is better for high-throughput analytics scenarios.

2. Use Kafka on Azure (Event Hubs with Kafka protocol): Event Hubs supports Kafka protocol, so this transport could optionally expose that interface for Kafka client compatibility.

3. Use Azure Event Grid: Event Grid is for event routing while Event Hubs is for event streaming. They complement each other for different patterns.

Affected Packages

• Encina (core)
• Encina.EntityFrameworkCore
• Encina.Dapper.*
• Encina.ADO.*
• Encina.AspNetCore
• Encina.OpenTelemetry
• Encina.Caching.*
• Other: Encina.AzureEventHubs (new package)

Additional Context

Acceptance Criteria

• Encina.AzureEventHubs package
• Producer implementation
• Consumer implementation with checkpointing
• Partition key support
• Consumer group support
• Batch operations
• Azure Schema Registry integration (optional)
• ServiceCollection extensions
• Unit tests
• Integration tests
• Documentation

Package Dependencies

• Azure.Messaging.EventHubs
• Azure.Messaging.EventHubs.Processor

Related Issues

• [FEATURE] Encina.GoogleCloudPubSub - Google Cloud Pub/Sub transport #237-[FEATURE] Encina.Dapr - Dapr PubSub integration #242 - Other transport implementations
• [FEATURE] ISchemaRegistryProvider: Abstraction and Implementations (Confluent, Azure, Apicurio) #602 - Schema Registry integration