Event Grid 및 이벤트 기반 아키텍처

이벤트 기반 아키텍처란 무엇인가요

Azure Event Grid란 무엇인가요

Event Grid 토픽 및 이벤트 구독

# Create a custom Event Grid topic
az eventgrid topic create \
  --name MyEventTopic \
  --resource-group MyRG \
  --location eastus

# Create a subscription routing to an Azure Function
az eventgrid event-subscription create \
  --name myFunctionSub \
  --source-resource-id '/subscriptions/.../providers/Microsoft.EventGrid/topics/MyEventTopic' \
  --endpoint '/subscriptions/.../providers/Microsoft.Web/sites/myFunctionApp/functions/EventHandler'

이벤트 스키마

// Event Grid event payload (Event Grid schema)
[
  {
    'id': 'b781910b-3000-4f19-a4c2-b6b9c4ca7a12',
    'eventType': 'Microsoft.Storage.BlobCreated',
    'subject': '/blobServices/default/containers/uploads/blobs/photo.jpg',
    'eventTime': '2025-01-15T12:30:00.000Z',
    'data': {
      'api': 'PutBlockList',
      'url': 'https://mystorageacct.blob.core.windows.net/uploads/photo.jpg',
      'contentType': 'image/jpeg',
      'contentLength': 524288
    },
    'dataVersion': '',
    'metadataVersion': '1'
  }
]

이벤트 필터링

# Subscribe to BlobCreated events for .jpg files only
az eventgrid event-subscription create \
  --name jpgImageSub \
  --source-resource-id '/subscriptions/.../storageAccounts/mystorageaccount' \
  --endpoint 'https://myfunction.azurewebsites.net/api/ProcessImage' \
  --included-event-types 'Microsoft.Storage.BlobCreated' \
  --subject-ends-with '.jpg'

Event Grid, Event Hubs 및 Service Bus 비교

// Decision guide:
// Event Grid  — 'Something happened, react to it'
//               Azure resource events, webhooks, low-latency fan-out
//               Price: per event (cheap for low volume)

// Event Hubs  — 'Capture a firehose of streaming data'
//               IoT telemetry, log aggregation, real-time analytics
//               Price: per throughput unit + capture

// Service Bus — 'Reliable message delivery between services'
//               Order processing, workflow steps, dead-letter queues
//               Price: per operation + messaging units

사용자 지정 이벤트 게시

# Get the topic endpoint and key
TOPIC_ENDPOINT=$(az eventgrid topic show --name MyEventTopic --resource-group MyRG --query endpoint -o tsv)
TOPIC_KEY=$(az eventgrid topic key list --name MyEventTopic --resource-group MyRG --query key1 -o tsv)

# Publish a custom event
curl -X POST $TOPIC_ENDPOINT \
  -H 'Content-Type: application/json' \
  -H "aeg-sas-key: $TOPIC_KEY" \
  -d '[{
    "id": "event-001",
    "eventType": "Contoso.OrderPlaced",
    "subject": "/orders/ORD-12345",
    "eventTime": "2025-01-15T12:00:00Z",
    "data": { "orderId": "ORD-12345", "total": 99.99 },
    "dataVersion": "1.0"
  }]'

배달 못한 편지 및 재시도 정책

# Configure dead-letter storage and retry for a subscription
az eventgrid event-subscription update \
  --name myFunctionSub \
  --source-resource-id '/subscriptions/.../topics/MyEventTopic' \
  --deadletter-endpoint '/subscriptions/.../storageAccounts/mystg/blobServices/default/containers/deadletter' \
  --max-delivery-attempts 30 \
  --event-ttl 1440  # 24 hours in minutes

Azure Event Hubs 개요

# Create an Event Hubs namespace and hub
az eventhubs namespace create \
  --name myEventHubNS \
  --resource-group MyRG \
  --location eastus \
  --sku Standard

az eventhubs eventhub create \
  --name telemetry \
  --namespace-name myEventHubNS \
  --resource-group MyRG \
  --partition-count 8 \
  --message-retention 3  # Days to retain events

안정적인 메시징을 위한 Azure Service Bus

# Create a Service Bus namespace and queue
az servicebus namespace create \
  --name myServiceBusNS \
  --resource-group MyRG \
  --location eastus \
  --sku Standard

az servicebus queue create \
  --name order-processing \
  --namespace-name myServiceBusNS \
  --resource-group MyRG \
  --enable-dead-lettering-on-message-expiration true \
  --max-delivery-count 10  # Move to dead-letter after 10 failed attempts

실용적인 이벤트 기반 패턴: 파일 처리

// Azure Function: process image on BlobCreated event
module.exports = async function (context, eventGridEvent) {
  const blobUrl = eventGridEvent.data.url;
  const blobName = eventGridEvent.subject.split('/blobs/').pop();

  context.log('Processing image:', blobName);

  // Process via Cognitive Services Vision API
  const tags = await analyzeImage(blobUrl);

  // Write metadata to Cosmos DB via output binding
  context.bindings.cosmosOutput = {
    id: blobName,
    tags,
    processedAt: new Date().toISOString()
  };

  context.log('Processing complete for:', blobName);
};

빠른 확인

단원 요약