azure-eventhub-ts – skill DevOps & Hạ tầng cho agent AI

Azure Event Hubs SDK for TypeScript

High-throughput event streaming and real-time data ingestion.

Installation

npm install @azure/event-hubs @azure/identity

For checkpointing with consumer groups:

npm install @azure/eventhubs-checkpointstore-blob @azure/storage-blob

Environment Variables

EVENTHUB_NAMESPACE=<namespace>.servicebus.windows.net
EVENTHUB_NAME=my-eventhub
STORAGE_ACCOUNT_NAME=<storage-account>
STORAGE_CONTAINER_NAME=checkpoints

Authentication

import { EventHubProducerClient, EventHubConsumerClient } from "@azure/event-hubs";
import { DefaultAzureCredential } from "@azure/identity";

const fullyQualifiedNamespace = process.env.EVENTHUB_NAMESPACE!;
const eventHubName = process.env.EVENTHUB_NAME!;
const credential = new DefaultAzureCredential();

// Producer
const producer = new EventHubProducerClient(fullyQualifiedNamespace, eventHubName, credential);

// Consumer
const consumer = new EventHubConsumerClient(
  "$Default", // Consumer group
  fullyQualifiedNamespace,
  eventHubName,
  credential
);

Core Workflow

Send Events

const producer = new EventHubProducerClient(namespace, eventHubName, credential);

// Create batch and add events
const batch = await producer.createBatch();
for (const body of [{ temperature: 72.5, deviceId: "sensor-1" }, { temperature: 68.2, deviceId: "sensor-2" }]) {
  if (!batch.tryAdd({ body })) throw new Error("Event does not fit this batch; flush/rebatch or reject oversized event");
}

await producer.sendBatch(batch);
await producer.close();

Send to Specific Partition

// By partition ID
const batch = await producer.createBatch({ partitionId: "0" });

// By partition key (consistent hashing)
const batch = await producer.createBatch({ partitionKey: "device-123" });

Receive Events (Simple)

const consumer = new EventHubConsumerClient("$Default", namespace, eventHubName, credential);

const subscription = consumer.subscribe({
  processEvents: async (events, context) => {
    for (const event of events) {
      console.log(`Partition: ${context.partitionId}, Body: ${JSON.stringify(event.body)}`);
    }
  },
  processError: async (err, context) => {
    console.error(`Error on partition ${context.partitionId}: ${err.message}`);
  },
});

// Stop after some time
setTimeout(async () => {
  await subscription.close();
  await consumer.close();
}, 60000);

Receive with Checkpointing (Production)

import { EventHubConsumerClient } from "@azure/event-hubs";
import { ContainerClient } from "@azure/storage-blob";
import { BlobCheckpointStore } from "@azure/eventhubs-checkpointstore-blob";

const containerClient = new ContainerClient(
  `https://${storageAccount}.blob.core.windows.net/${containerName}`,
  credential
);

const checkpointStore = new BlobCheckpointStore(containerClient);

const consumer = new EventHubConsumerClient(
  "$Default",
  namespace,
  eventHubName,
  credential,
  checkpointStore
);

const subscription = consumer.subscribe({
  processEvents: async (events, context) => {
    for (const event of events) {
      console.log(`Processing: ${JSON.stringify(event.body)}`);
    }
    // Checkpoint after processing batch
    if (events.length > 0) {
      await context.updateCheckpoint(events[events.length - 1]);
    }
  },
  processError: async (err, context) => {
    console.error(`Error: ${err.message}`);
  },
});

Receive from Specific Position

const subscription = consumer.subscribe({
  processEvents: async (events, context) => { /* ... */ },
  processError: async (err, context) => { /* ... */ },
}, {
  startPosition: {
    // Start from beginning
    "0": { offset: "@earliest" },
    // Start from end (new events only)
    "1": { offset: "@latest" },
    // Start from specific offset
    "2": { offset: "12345" },
    // Start from specific time
    "3": { enqueuedOn: new Date("2024-01-01") },
  },
});

Event Hub Properties

// Get hub info
const hubProperties = await producer.getEventHubProperties();
console.log(`Partitions: ${hubProperties.partitionIds}`);

// Get partition info
const partitionProperties = await producer.getPartitionProperties("0");
console.log(`Last sequence: ${partitionProperties.lastEnqueuedSequenceNumber}`);

Batch Processing Options

const subscription = consumer.subscribe(
  {
    processEvents: async (events, context) => { /* ... */ },
    processError: async (err, context) => { /* ... */ },
  },
  {
    maxBatchSize: 100,           // Max events per batch
    maxWaitTimeInSeconds: 30,    // Max wait for batch
  }
);

Key Types

import {
  EventHubProducerClient,
  EventHubConsumerClient,
  EventData,
  ReceivedEventData,
  PartitionContext,
  Subscription,
  SubscriptionEventHandlers,
  CreateBatchOptions,
  EventPosition,
} from "@azure/event-hubs";

import { BlobCheckpointStore } from "@azure/eventhubs-checkpointstore-blob";

Event Properties

// Send with properties
const batch = await producer.createBatch();
batch.tryAdd({
  body: { data: "payload" },
  properties: {
    eventType: "telemetry",
    deviceId: "sensor-1",
  },
  contentType: "application/json",
  correlationId: "request-123",
});

// Access in receiver
consumer.subscribe({
  processEvents: async (events, context) => {
    for (const event of events) {
      console.log(`Type: ${event.properties?.eventType}`);
      console.log(`Sequence: ${event.sequenceNumber}`);
      console.log(`Enqueued: ${event.enqueuedTimeUtc}`);
      console.log(`Offset: ${event.offset}`);
    }
  },
});

Error Handling

consumer.subscribe({
  processEvents: async (events, context) => {
    try {
      for (const event of events) {
        await processEvent(event);
      }
      if (events.length > 0) await context.updateCheckpoint(events[events.length - 1]);
    } catch (error) {
      // Don't checkpoint on error - events will be reprocessed
      console.error("Processing failed:", error);
    }
  },
  processError: async (err, context) => {
    if (err.name === "MessagingError") {
      // Transient error - SDK will retry
      console.warn("Transient error:", err.message);
    } else {
      // Fatal error
      console.error("Fatal error:", err);
    }
  },
});

Best Practices

1. Use checkpointing - Checkpoint only after durable processing; delivery can repeat, so make handlers idempotent

2. Batch sends - Use createBatch() for efficient sending

3. Partition keys - Use partition keys to ensure ordering for related events

4. Consumer groups - Use separate consumer groups for different processing pipelines

5. Handle errors gracefully - Don't checkpoint on processing failures

6. Close clients - Always close producer/consumer when done

7. Monitor lag - Track lastEnqueuedSequenceNumber vs processed sequence

When to Use

This skill is applicable to execute the workflow or actions described in the overview.

Limitations

• Use this skill only when the task clearly matches the scope described above.
• Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
• Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.

Checkpointing records progress; it does not make external side effects exactly once or guarantee an immediate retry when an error is swallowed. Use stable event IDs, durable deduplication and an explicit retry/dead-letter design. Inspect retryability rather than treating every MessagingError as transient. Examples are fragments; supply imports/handlers and redact event bodies. [Official SDK guide](https://learn.microsoft.com/en-us/javascript/api/overview/azure/event-hubs-readme?view=azure-node-latest).