azure-eventhub-py

作者: microsoft

Azure Event Hubs SDK for Python streaming. Use for high-throughput event ingestion, producers, consumers, and checkpointing. Triggers: "event hubs", "EventHubProducerClient", "EventHubConsumerClient", "streaming", "partitions".

npx skills add https://github.com/microsoft/skills --skill azure-eventhub-py

Azure Event Hubs SDK for Python

Big data streaming platform for high-throughput event ingestion.

Installation

pip install azure-eventhub azure-identity
# For checkpointing with blob storage
pip install azure-eventhub-checkpointstoreblob-aio

Environment Variables

EVENT_HUB_FULLY_QUALIFIED_NAMESPACE=<namespace>.servicebus.windows.net  # Required for all auth methods
EVENT_HUB_NAME=my-eventhub  # Required for all auth methods
STORAGE_ACCOUNT_URL=https://<account>.blob.core.windows.net  # Required for checkpoint storage
CHECKPOINT_CONTAINER=checkpoints  # Required for checkpoint storage
AZURE_TOKEN_CREDENTIALS=prod # Required only if DefaultAzureCredential is used in production

Authentication & Lifecycle

🔑 Two rules apply to every code sample below:

  1. Prefer DefaultAzureCredential. It works locally (Azure CLI / VS Code / Developer CLI) and in Azure (managed identity, workload identity) with no code change. Avoid connection strings, account/API keys — they bypass Entra audit and rotation.
    • Local dev: DefaultAzureCredential works as-is.
    • Production: set AZURE_TOKEN_CREDENTIALS=prod (or AZURE_TOKEN_CREDENTIALS=<specific_credential>) to constrain the credential chain to production-safe credentials.
  2. Wrap every client in a context manager so HTTP transports, sockets, and token caches are released deterministically:
    • Sync: with <Client>(...) as client:
    • Async: async with <Client>(...) as client: and async with DefaultAzureCredential() as credential: (from azure.identity.aio)

Snippets may abbreviate this setup, but production code should always follow both rules.

from azure.identity import DefaultAzureCredential, ManagedIdentityCredential
from azure.eventhub import EventHubProducerClient, EventHubConsumerClient

# Local dev: DefaultAzureCredential. Production: set AZURE_TOKEN_CREDENTIALS=prod or AZURE_TOKEN_CREDENTIALS=<specific_credential>
credential = DefaultAzureCredential(require_envvar=True)
# Or use a specific credential directly in production:
# See https://learn.microsoft.com/python/api/overview/azure/identity-readme?view=azure-python#credential-classes
# credential = ManagedIdentityCredential()
namespace = "<namespace>.servicebus.windows.net"
eventhub_name = "my-eventhub"

# Producer
with EventHubProducerClient(
    fully_qualified_namespace=namespace,
    eventhub_name=eventhub_name,
    credential=credential
) as producer:
    # Use producer here (see following sections for operations)
    ...

# Consumer
with EventHubConsumerClient(
    fully_qualified_namespace=namespace,
    eventhub_name=eventhub_name,
    consumer_group="$Default",
    credential=credential
) as consumer:
    # Use consumer here (see following sections for operations)
    ...

Client Types

ClientPurpose
EventHubProducerClientSend events to Event Hub
EventHubConsumerClientReceive events from Event Hub
BlobCheckpointStoreTrack consumer progress

Send Events

from azure.eventhub import EventHubProducerClient, EventData
from azure.identity import DefaultAzureCredential

with EventHubProducerClient(
    fully_qualified_namespace="<namespace>.servicebus.windows.net",
    eventhub_name="my-eventhub",
    credential=DefaultAzureCredential()
) as producer:
    # Create batch (handles size limits)
    event_data_batch = producer.create_batch()
    
    for i in range(10):
        try:
            event_data_batch.add(EventData(f"Event {i}"))
        except ValueError:
            # Batch is full, send and create new one
            producer.send_batch(event_data_batch)
            event_data_batch = producer.create_batch()
            event_data_batch.add(EventData(f"Event {i}"))
    
    # Send remaining
    producer.send_batch(event_data_batch)

Send to Specific Partition

# By partition ID
event_data_batch = producer.create_batch(partition_id="0")

# By partition key (consistent hashing)
event_data_batch = producer.create_batch(partition_key="user-123")

Receive Events

Simple Receive

from azure.eventhub import EventHubConsumerClient

def on_event(partition_context, event):
    print(f"Partition: {partition_context.partition_id}")
    print(f"Data: {event.body_as_str()}")
    partition_context.update_checkpoint(event)

with EventHubConsumerClient(
    fully_qualified_namespace="<namespace>.servicebus.windows.net",
    eventhub_name="my-eventhub",
    consumer_group="$Default",
    credential=DefaultAzureCredential()
) as consumer:
    consumer.receive(
        on_event=on_event,
        starting_position="-1",  # Beginning of stream
    )

With Blob Checkpoint Store (Production)

from azure.eventhub import EventHubConsumerClient
from azure.eventhub.extensions.checkpointstoreblob import BlobCheckpointStore
from azure.identity import DefaultAzureCredential

checkpoint_store = BlobCheckpointStore(
    blob_account_url="https://<account>.blob.core.windows.net",
    container_name="checkpoints",
    credential=DefaultAzureCredential()
)

with EventHubConsumerClient(
    fully_qualified_namespace="<namespace>.servicebus.windows.net",
    eventhub_name="my-eventhub",
    consumer_group="$Default",
    credential=DefaultAzureCredential(),
    checkpoint_store=checkpoint_store
) as consumer:
    def on_event(partition_context, event):
        print(f"Received: {event.body_as_str()}")
        # Checkpoint after processing
        partition_context.update_checkpoint(event)

    consumer.receive(on_event=on_event)

Async Client

from azure.eventhub.aio import EventHubProducerClient, EventHubConsumerClient
from azure.identity.aio import DefaultAzureCredential
import asyncio

async def send_events():
    credential = DefaultAzureCredential()
    
    async with EventHubProducerClient(
        fully_qualified_namespace="<namespace>.servicebus.windows.net",
        eventhub_name="my-eventhub",
        credential=credential
    ) as producer:
        batch = await producer.create_batch()
        batch.add(EventData("Async event"))
        await producer.send_batch(batch)

async def receive_events():
    async def on_event(partition_context, event):
        print(event.body_as_str())
        await partition_context.update_checkpoint(event)
    
    async with EventHubConsumerClient(
        fully_qualified_namespace="<namespace>.servicebus.windows.net",
        eventhub_name="my-eventhub",
        consumer_group="$Default",
        credential=DefaultAzureCredential()
    ) as consumer:
        await consumer.receive(on_event=on_event)

asyncio.run(send_events())

Event Properties

event = EventData("My event body")

# Set properties
event.properties = {"custom_property": "value"}
event.content_type = "application/json"

# Read properties (on receive)
print(event.body_as_str())
print(event.sequence_number)
print(event.offset)
print(event.enqueued_time)
print(event.partition_key)

Get Event Hub Info

with producer:
    info = producer.get_eventhub_properties()
    print(f"Name: {info['name']}")
    print(f"Partitions: {info['partition_ids']}")
    
    for partition_id in info['partition_ids']:
        partition_info = producer.get_partition_properties(partition_id)
        print(f"Partition {partition_id}: {partition_info['last_enqueued_sequence_number']}")

Best Practices

  1. Pick sync OR async and stay consistent. Do not mix azure.xxx sync clients with azure.xxx.aio async clients in the same call path. Choose one mode per module.
  2. Always use context managers for clients and async credentials. Wrap every client in with Client(...) as client: (sync) or async with Client(...) as client: (async) for proper cleanup. For async DefaultAzureCredential from azure.identity.aio, also use async with credential: so tokens and transports are cleaned up.
  3. Use DefaultAzureCredential for portable auth across local dev and Azure (avoid connection strings / API keys when possible).
  4. Use batches for sending multiple events
  5. Use checkpoint store in production for reliable processing
  6. Use async client for high-throughput scenarios
  7. Use partition keys for ordered delivery within a partition
  8. Handle batch size limits — catch ValueError when batch is full
  9. Set appropriate consumer groups for different applications

Reference Files

FileContents
references/checkpointing.mdCheckpoint store patterns, blob checkpointing, checkpoint strategies
references/partitions.mdPartition management, load balancing, starting positions
scripts/setup_consumer.pyCLI for Event Hub info, consumer setup, and event sending/receiving

來自 microsoft 的更多技能

oss-growth
microsoft
開源增長駭客角色
agent-framework-azure-ai-py
microsoft
使用Microsoft Agent Framework Python SDK(agent-framework-azure-ai)构建Azure AI Foundry代理。适用于使用AzureAIAgentsProvider创建持久化代理、使用托管工具(代码解释器、文件搜索、网络搜索)、集成MCP服务器、管理对话线程或实现流式响应。涵盖函数工具、结构化输出和多工具代理。
development
airunway-aks-setup
microsoft
在AKS上設定AI Runway——從裸叢集到執行模型。涵蓋叢集驗證、控制器安裝、GPU評估、供應商設定及首次部署。時機:「設定AI Runway」、「上線AKS叢集」、「安裝AI Runway」、「airunway設定」、「部署模型至AKS」、「在AKS上進行GPU推論」、「在AKS上設定KAITO」、「在AKS上執行LLM」、「在AKS上使用vLLM」、「在AKS上設定模型服務」、「AI Runway控制器」。
devops
appinsights-instrumentation
microsoft
使用Azure Application Insights檢測Web應用程式的指南。提供遙測模式、SDK設定與組態參考。適用時機:如何檢測應用程式、App Insights SDK、遙測模式、什麼是App Insights、Application Insights指南、檢測範例、APM最佳實踐。
devops
applicationinsights-web-ts
microsoft
使用Application Insights JavaScript SDK(@microsoft/applicationinsights-web)為瀏覽器/Web應用程式進行檢測。適用於真實使用者監控(RUM)——頁面檢視、點擊、AJAX/fetch依賴、例外、自訂事件,以及與後端OpenTelemetry追蹤關聯的瀏覽器端GenAI代理追蹤。涵蓋SDK載入器指令碼與npm設定、框架擴充(React、React Native、Angular)、點擊分析、遙測初始化器,以及從瀏覽器發出的代理/工具/模型span的OTel GenAI語意慣例。
devops
azure-ai-anomalydetector-java
microsoft
使用適用於 Java 的 Azure AI 異常偵測器 SDK 建置異常偵測應用程式。在實作單變量/多變量異常偵測、時間序列分析或 AI 驅動監控時使用。
development
azure-ai-language-conversations-py
microsoft
使用 azure-ai-language-conversations Python SDK 實作對話語言理解(CLU)。當使用 ConversationAnalysisClient 分析對話意圖與實體、建置 NLP 功能,或將語言理解整合至應用程式時使用。
development
azure-ai-ml-py
microsoft
Azure Machine Learning SDK v2 for Python。用於機器學習工作區、作業、模型、資料集、計算資源與管線。 觸發詞:「azure-ai-ml」、「MLClient」、「workspace」、「model registry」、「training jobs」、「datasets」。
development