In brief
The SDK Azure Event Hubs for Java builds real-time streaming applications: broadband ingestion and event architecture based on a partitioned journal. The skill details customer creation, consumption and emission patterns, positions in flow (event positions), error management, and resource cleaning. This is not Event Grid (notifications) or Bus Service (files and topics). Authentication goes through the variables and credentials of the guide; no connection strings are owned by the repository. Good skill practices (client closure, take-back positions) are part of the contract of use.
What it does
- Covers production and consumption of an Event Hubs feed, with event positions to resume playback.
- Emphasises on resource cleaning and explicit error processing.
- Do not replace a task line or a notification bus.
- Authenticated by Entra rather than by a namespace key written in the code.
Who it is for, and an example
Java teams that abound or produce a bulky flow, with a reliable recovery on partitions.
Example request: A consumer returns to a known position, treats the lot, then closes the customer properly, authenticated by managed identity.
When implementing Event Hubs streaming, broadband ingestion or event architecture in Java. Example written by iadujour.fr from the SKILL.md, not tested.
Install
npx skills add microsoft/skills (selection of the skill in the assistant, README).
npx skills add microsoft/skillsWorks with
- GitHub Copilot
- Copilot CLI
Source: README: examples of Copilot CLI and GitHub Copilot in VS Code; installation `npx skills add` sends files to the chosen agent directory, without an exhaustive list.
Level: Intermediate
Suppose a publisher's project and mastery of the corresponding SDK or CLI.
Good to know
SKILL.md excerpt
The original description, which the agent reads to decide whether to load the skill (in English):
Build real-time streaming applications with Azure Event Hubs SDK for Java. Use when implementing event streaming, high-throughput data ingestion, or building event-driven architectures.
Quoted from SKILL.md at commit e7ee58f . The folder also contains 1 extra file (references/).
The SKILL.md has 362 lines and 23 main sections:
- Installation
- Client Creation
- EventHubProducerClient
- With DefaultAzureCredential
- EventHubConsumerClient
- Async Clients
- Core Patterns
- Send Single Event
- Send Event Batch
- Send to Specific Partition
- Send with Partition Key
- Event with Properties
- Receive Events (Simple)
- EventProcessorClient (Production)
- Batch Processing
- Async Receiving
- Get Event Hub Properties
- Event Positions
- Error Handling
- Resource Cleanup
- Environment Variables
- Best Practices
- Trigger Phrases
Licence: MIT
Checked on 10 November 2026 : repository LICENSE file . Licence declared in the frontmatter: “MIT”. Rights holder: Copyright (c) Microsoft Corporation..
The licence allows reuse: the full SKILL.md text is reproduced below, unmodified, with its licence notice.
Full SKILL.md text
Reproduced unmodified from microsoft/skills (commit e7ee58f) , © Microsoft, under licence MIT (original licence file ).
Show the original SKILL.md (362 lines, in English)
---
name: azure-eventhub-java
description: Build real-time streaming applications with Azure Event Hubs SDK for Java. Use when implementing event streaming, high-throughput data ingestion, or building event-driven architectures.
license: MIT
metadata:
author: Microsoft
version: "1.0.0"
package: com.azure:azure-messaging-eventhubs
---
# Azure Event Hubs SDK for Java
Build real-time streaming applications using the Azure Event Hubs SDK for Java.
## Installation
```xml
<dependency>
<groupId>com.azure</groupId>
<artifactId>azure-messaging-eventhubs</artifactId>
<version>5.19.0</version>
</dependency>
<!-- For checkpoint store (production) -->
<dependency>
<groupId>com.azure</groupId>
<artifactId>azure-messaging-eventhubs-checkpointstore-blob</artifactId>
<version>1.20.0</version>
</dependency>
```
## Client Creation
### EventHubProducerClient
```java
import com.azure.messaging.eventhubs.EventHubProducerClient;
import com.azure.messaging.eventhubs.EventHubClientBuilder;
// With connection string
EventHubProducerClient producer = new EventHubClientBuilder()
.connectionString("<connection-string>", "<event-hub-name>")
.buildProducerClient();
// Full connection string with EntityPath
EventHubProducerClient producer = new EventHubClientBuilder()
.connectionString("<connection-string-with-entity-path>")
.buildProducerClient();
```
### With DefaultAzureCredential
```java
import com.azure.core.credential.TokenCredential;
import com.azure.identity.AzureIdentityEnvVars;
import com.azure.identity.DefaultAzureCredentialBuilder;
import com.azure.identity.ManagedIdentityCredentialBuilder;
// Local dev: DefaultAzureCredential. Production: set AZURE_TOKEN_CREDENTIALS=prod or AZURE_TOKEN_CREDENTIALS=<specific_credential>
TokenCredential credential = new DefaultAzureCredentialBuilder()
.requireEnvVars(AzureIdentityEnvVars.AZURE_TOKEN_CREDENTIALS)
.build();
// Or use a specific credential directly in production:
// See https://learn.microsoft.com/java/api/overview/azure/identity-readme?view=azure-java-stable#credential-classes
// TokenCredential credential = new ManagedIdentityCredentialBuilder().build();
EventHubProducerClient producer = new EventHubClientBuilder()
.fullyQualifiedNamespace("<namespace>.servicebus.windows.net")
.eventHubName("<event-hub-name>")
.credential(credential)
.buildProducerClient();
```
### EventHubConsumerClient
```java
import com.azure.messaging.eventhubs.EventHubConsumerClient;
EventHubConsumerClient consumer = new EventHubClientBuilder()
.connectionString("<connection-string>", "<event-hub-name>")
.consumerGroup(EventHubClientBuilder.DEFAULT_CONSUMER_GROUP_NAME)
.buildConsumerClient();
```
### Async Clients
```java
import com.azure.messaging.eventhubs.EventHubProducerAsyncClient;
import com.azure.messaging.eventhubs.EventHubConsumerAsyncClient;
EventHubProducerAsyncClient asyncProducer = new EventHubClientBuilder()
.connectionString("<connection-string>", "<event-hub-name>")
.buildAsyncProducerClient();
EventHubConsumerAsyncClient asyncConsumer = new EventHubClientBuilder()
.connectionString("<connection-string>", "<event-hub-name>")
.consumerGroup("$Default")
.buildAsyncConsumerClient();
```
## Core Patterns
### Send Single Event
```java
import com.azure.messaging.eventhubs.EventData;
EventData eventData = new EventData("Hello, Event Hubs!");
producer.send(Collections.singletonList(eventData));
```
### Send Event Batch
```java
import com.azure.messaging.eventhubs.EventDataBatch;
import com.azure.messaging.eventhubs.models.CreateBatchOptions;
// Create batch
EventDataBatch batch = producer.createBatch();
// Add events (returns false if batch is full)
for (int i = 0; i < 100; i++) {
EventData event = new EventData("Event " + i);
if (!batch.tryAdd(event)) {
// Batch is full, send and create new batch
producer.send(batch);
batch = producer.createBatch();
batch.tryAdd(event);
}
}
// Send remaining events
if (batch.getCount() > 0) {
producer.send(batch);
}
```
### Send to Specific Partition
```java
CreateBatchOptions options = new CreateBatchOptions()
.setPartitionId("0");
EventDataBatch batch = producer.createBatch(options);
batch.tryAdd(new EventData("Partition 0 event"));
producer.send(batch);
```
### Send with Partition Key
```java
CreateBatchOptions options = new CreateBatchOptions()
.setPartitionKey("customer-123");
EventDataBatch batch = producer.createBatch(options);
batch.tryAdd(new EventData("Customer event"));
producer.send(batch);
```
### Event with Properties
```java
EventData event = new EventData("Order created");
event.getProperties().put("orderId", "ORD-123");
event.getProperties().put("customerId", "CUST-456");
event.getProperties().put("priority", 1);
producer.send(Collections.singletonList(event));
```
### Receive Events (Simple)
```java
import com.azure.messaging.eventhubs.models.EventPosition;
import com.azure.messaging.eventhubs.models.PartitionEvent;
// Receive from specific partition
Iterable<PartitionEvent> events = consumer.receiveFromPartition(
"0", // partitionId
10, // maxEvents
EventPosition.earliest(), // startingPosition
Duration.ofSeconds(30) // timeout
);
for (PartitionEvent partitionEvent : events) {
EventData event = partitionEvent.getData();
System.out.println("Body: " + event.getBodyAsString());
System.out.println("Sequence: " + event.getSequenceNumber());
System.out.println("Offset: " + event.getOffset());
}
```
### EventProcessorClient (Production)
```java
import com.azure.messaging.eventhubs.EventProcessorClient;
import com.azure.messaging.eventhubs.EventProcessorClientBuilder;
import com.azure.messaging.eventhubs.checkpointstore.blob.BlobCheckpointStore;
import com.azure.storage.blob.BlobContainerAsyncClient;
import com.azure.storage.blob.BlobContainerClientBuilder;
// Create checkpoint store
BlobContainerAsyncClient blobClient = new BlobContainerClientBuilder()
.connectionString("<storage-connection-string>")
.containerName("checkpoints")
.buildAsyncClient();
// Create processor
EventProcessorClient processor = new EventProcessorClientBuilder()
.connectionString("<eventhub-connection-string>", "<event-hub-name>")
.consumerGroup("$Default")
.checkpointStore(new BlobCheckpointStore(blobClient))
.processEvent(eventContext -> {
EventData event = eventContext.getEventData();
System.out.println("Processing: " + event.getBodyAsString());
// Checkpoint after processing
eventContext.updateCheckpoint();
})
.processError(errorContext -> {
System.err.println("Error: " + errorContext.getThrowable().getMessage());
System.err.println("Partition: " + errorContext.getPartitionContext().getPartitionId());
})
.buildEventProcessorClient();
// Start processing
processor.start();
// Keep running...
Thread.sleep(Duration.ofMinutes(5).toMillis());
// Stop gracefully
processor.stop();
```
### Batch Processing
```java
EventProcessorClient processor = new EventProcessorClientBuilder()
.connectionString("<connection-string>", "<event-hub-name>")
.consumerGroup("$Default")
.checkpointStore(new BlobCheckpointStore(blobClient))
.processEventBatch(eventBatchContext -> {
List<EventData> events = eventBatchContext.getEvents();
System.out.printf("Received %d events%n", events.size());
for (EventData event : events) {
// Process each event
System.out.println(event.getBodyAsString());
}
// Checkpoint after batch
eventBatchContext.updateCheckpoint();
}, 50) // maxBatchSize
.processError(errorContext -> {
System.err.println("Error: " + errorContext.getThrowable());
})
.buildEventProcessorClient();
```
### Async Receiving
```java
asyncConsumer.receiveFromPartition("0", EventPosition.latest())
.subscribe(
partitionEvent -> {
EventData event = partitionEvent.getData();
System.out.println("Received: " + event.getBodyAsString());
},
error -> System.err.println("Error: " + error),
() -> System.out.println("Complete")
);
```
### Get Event Hub Properties
```java
// Get hub info
EventHubProperties hubProps = producer.getEventHubProperties();
System.out.println("Hub: " + hubProps.getName());
System.out.println("Partitions: " + hubProps.getPartitionIds());
// Get partition info
PartitionProperties partitionProps = producer.getPartitionProperties("0");
System.out.println("Begin sequence: " + partitionProps.getBeginningSequenceNumber());
System.out.println("Last sequence: " + partitionProps.getLastEnqueuedSequenceNumber());
System.out.println("Last offset: " + partitionProps.getLastEnqueuedOffset());
```
## Event Positions
```java
// Start from beginning
EventPosition.earliest()
// Start from end (new events only)
EventPosition.latest()
// From specific offset
EventPosition.fromOffset(12345L)
// From specific sequence number
EventPosition.fromSequenceNumber(100L)
// From specific time
EventPosition.fromEnqueuedTime(Instant.now().minus(Duration.ofHours(1)))
```
## Error Handling
```java
import com.azure.messaging.eventhubs.models.ErrorContext;
.processError(errorContext -> {
Throwable error = errorContext.getThrowable();
String partitionId = errorContext.getPartitionContext().getPartitionId();
if (error instanceof AmqpException) {
AmqpException amqpError = (AmqpException) error;
if (amqpError.isTransient()) {
System.out.println("Transient error, will retry");
}
}
System.err.printf("Error on partition %s: %s%n", partitionId, error.getMessage());
})
```
## Resource Cleanup
```java
// Always close clients
try {
producer.send(batch);
} finally {
producer.close();
}
// Or use try-with-resources
try (EventHubProducerClient producer = new EventHubClientBuilder()
.connectionString(connectionString, eventHubName)
.buildProducerClient()) {
producer.send(events);
}
```
## Environment Variables
```bash
EVENT_HUBS_CONNECTION_STRING=Endpoint=sb://<namespace>.servicebus.windows.net/;SharedAccessKeyName=... # Alternative to Entra ID auth
EVENT_HUBS_NAME=<event-hub-name> # Required for event hub name
STORAGE_CONNECTION_STRING=<for-checkpointing> # Alternative to Entra ID auth for checkpointing
AZURE_TOKEN_CREDENTIALS=prod # Required only if DefaultAzureCredential is used in production
```
## Best Practices
1. **Use EventProcessorClient**: For production, provides load balancing and checkpointing
2. **Batch Events**: Use `EventDataBatch` for efficient sending
3. **Partition Keys**: Use for ordering guarantees within a partition
4. **Checkpointing**: Checkpoint after processing to avoid reprocessing
5. **Error Handling**: Handle transient errors with retries
6. **Close Clients**: Always close producer/consumer when done
## Trigger Phrases
- "Event Hubs Java"
- "event streaming Azure"
- "real-time data ingestion"
- "EventProcessorClient"
- "event hub producer consumer"
- "partition processing"
Comments.
A field note, a question, a precision on this skill? Comments are reviewed before publication.
Published comments
Loading comments…
Leave a comment
Moderation charter. Every comment is reviewed before publication. We refuse hate speech, insulting or discriminatory language, advertising, spam and promotional links, off-topic posts and other people’s personal data.