Serverless AWS Lambda Development · レッスン

失敗時のデッドレターキュー（DLQ）

SQSまたはSNSでデッドレターキュー（DLQ）を設定し、失敗した非同期Lambda呼び出しを捕捉・処理して、システムの回復力とデバッグ性を高めます。

「失敗時のデッドレターキュー（DLQ）」はCoddyKit上の無料Serverless AWS Lambda Developmentレッスンです。これはレッスン2/4です。下記で完全なレッスンを無料で読むことができます。その後、ブラウザ内の組み込みコードエディタと24時間対応のAIチューターでハンズオン演習できます。これはServerless AWS Lambda Development学習パスの一部であり、ウェブとCoddyKitアプリ全体で進捗が同期されます。 Serverless AWS Lambda Developmentコースには全4レッスンが含まれています。

このレッスンの一部はまだ翻訳されておらず、英語で表示されています。

Why Dead Letter Queues?

When building serverless applications, especially with asynchronous Lambda functions, what happens if an invocation fails repeatedly?

Without a proper mechanism, these failed events might simply be discarded, leading to data loss or unaddressed issues. This is where Dead Letter Queues (DLQs) come in.

Async Lambda Invocation Review

First, let's quickly recap how asynchronous Lambda invocations work. When you invoke a Lambda function asynchronously (e.g., via S3, SNS, or direct API call with InvocationType: Event):

Lambda places the event in an internal queue.
It then attempts to invoke your function.
If the function fails, Lambda automatically retries the invocation up to two times.

Unhandled Async Failures

What happens if your Lambda function still fails after all automatic retries (initial attempt + two retries)?

By default, if no DLQ is configured, the event is simply discarded. This means you lose valuable information about what went wrong and the event data itself, making debugging and error recovery difficult.

Dead Letter Queue Defined

A Dead Letter Queue (DLQ) is a destination for events that Lambda couldn't successfully process after exhausting all retry attempts.

Think of it as a 'parking lot' for problematic messages. Instead of disappearing, these failed events are sent to your chosen DLQ destination, allowing you to inspect, debug, and potentially re-process them later.

DLQ Destinations: SQS or SNS?

You can configure two types of AWS services as DLQ destinations for your Lambda functions:

Amazon SQS (Simple Queue Service): A message queue. Failed events are sent to the SQS queue, where they await processing. This is a pull-based model.
Amazon SNS (Simple Notification Service): A topic. Failed events are published to an SNS topic, which can then notify subscribers (e.g., email, other Lambda functions). This is a push-based model.

SQS is generally preferred for re-processing, while SNS is good for immediate notifications.

Configuring an SQS DLQ

To use an SQS queue as a DLQ, you first need to create one. It's a standard SQS queue, but often named to indicate its purpose (e.g., my-lambda-dlq).

Here's how you might create a standard SQS queue using the AWS CLI:

aws sqs create-queue \
  --queue-name my-lambda-dlq

Connect Lambda to DLQ

Once your SQS queue is ready, you configure your Lambda function to use it as its DLQ. This involves updating the function's configuration.

You also need to ensure your Lambda's IAM execution role has permissions to send messages to the SQS queue (sqs:SendMessage).

aws lambda update-function-configuration \
  --function-name MyFailingLambda \
  --dead-letter-config TargetArn=arn:aws:sqs:REGION:ACCOUNT_ID:my-lambda-dlq

Demo: Lambda Failure to DLQ

Consider this Python Lambda function. It processes an event, but if the event contains "should_fail": true, it will raise an exception.

When invoked asynchronously, after retries, an event causing this failure would be sent to the configured DLQ.

def lambda_handler(event, context):
    print(f"Processing event: {event}")
    # Simulate an error condition
    if event.get("should_fail", False):
        raise Exception("Simulated processing error!")
    return {
        'statusCode': 200,
        'body': 'Processed successfully!'
    }

# This part makes it runnable outside Lambda for demonstration
if __name__ == "__main__":
    print("--- Simulating a successful invocation ---")
    result_success = lambda_handler({"key": "value"}, None)
    print(f"Success Result: {result_success}\n")

    print("--- Simulating a failed invocation ---")
    try:
        result_fail = lambda_handler({"should_fail": True}, None)
        print(f"Failure Result: {result_fail}")
    except Exception as e:
        print(f"Caught expected error: {e}")
        print("This event would eventually go to a DLQ after retries.")

Managing Failed Events

Once events are in your DLQ, you can:

Monitor: Use Amazon CloudWatch to track the number of messages in the DLQ.
Inspect: View the content of the messages to understand the failure.
Re-process: Move messages back to the original queue or trigger manual processing once the underlying issue is resolved.

This provides a crucial safety net for your asynchronous workflows.

DLQ Quick Check

You've learned about Dead Letter Queues and their importance. Let's test your understanding!

Recap: DLQs for Resilience

In this lesson, we explored Dead Letter Queues (DLQs) and their role in building resilient serverless applications. You learned:

DLQs prevent data loss from failed asynchronous Lambda invocations.
AWS SQS and SNS can serve as DLQ destinations.
How to configure a Lambda function with a DLQ.
The importance of monitoring and managing events in your DLQ.

DLQs are essential for robust error handling in event-driven architectures.

よくある質問

「失敗時のデッドレターキュー（DLQ）」レッスンは無料ですか？

はい。「失敗時のデッドレターキュー（DLQ）」の完全なテキストはこのウェブで無料で読めます。インタラクティブに演習し（組み込みコードエディタと24時間対応のAIチューター）、Serverless AWS Lambda Developmentコースの残りをアンロックするには、CoddyKit PROにアップグレードしてください。 Serverless AWS Lambda Developmentコースには全4レッスンが含まれています。

「失敗時のデッドレターキュー（DLQ）」で何を学びますか？

SQSまたはSNSでデッドレターキュー（DLQ）を設定し、失敗した非同期Lambda呼び出しを捕捉・処理して、システムの回復力とデバッグ性を高めます。ブラウザで直接実行するハンズオンコードでServerless AWS Lambda Developmentを演習し、24時間対応のAIチューターがレッスンを進める中での質問に答えます。

Serverless AWS Lambda Developmentを始めるのに経験は必要ですか？

事前経験は必要ありません。CoddyKitのServerless AWS Lambda Developmentは初級者から上級者向けに構成されているため、ここから始めるか最初から始めて、自分のペースで進むことができます。これはレッスン2/4です。

「失敗時のデッドレターキュー（DLQ）」レッスンにはどのくらい時間がかかりますか？

ほとんどのCoddyKitレッスンは約5～10分かかります。各レッスンはコンパクトでインタラクティブなので、着実に進歩し、ウェブとアプリ全体で正確に前回の場所から再開できます。

このServerless AWS Lambda Developmentレッスンでコードを書いて実行できますか？

はい。すべてのServerless AWS Lambda Developmentレッスンに組み込みコードエディタが含まれているため、ブラウザでリアルコードを書いて実行し、即座のAIフィードバックを取得できます。ローカル設定は不要です。

このコースのすべてのレッスン

← Serverless AWS Lambda Developmentに戻る