Capturing records of Lambda asynchronous invocations

Lambda can send records of asynchronous invocations to one of the following AWS services.

The invocation record contains details about the request and response in JSON format. You can configure separate destinations for events that are processed successfully, and events that fail all processing attempts. Alternatively, you can configure a standard Amazon SQS queue or standard Amazon SNS topic as a dead-letter queue for discarded events. For dead-letter queues, Lambda only sends the content of the event, without details about the response.

If Lambda can't send a record to a destination you have configured, it sends a DestinationDeliveryFailures metric to Amazon CloudWatch. This can happen if your configuration includes an unsupported destination type, such as an Amazon SQS FIFO queue or an Amazon SNS FIFO topic. Delivery errors can also occur due to permissions errors and size limits. For more information on Lambda invocation metrics, see Invocation metrics.

Adding a destination

To retain records of asynchronous invocations, add a destination to your function. You can choose to send either successful or failed invocations to a destination. Each function can have multiple destinations, so you can configure separate destinations for successful and failed events. Each record sent to the destination is a JSON document with details about the invocation. Like error handling settings, you can configure destinations on a function, function version, or alias.

The following table lists supported destinations for asynchronous invocation records. For Lambda to successfully send records to your chosen destination, ensure that your function's execution role also contains the relevant permissions. The table also describes how each destination type receives the JSON invocation record.

The following steps describe how to configure a destination for a function using the Lambda console and the AWS CLI.

Security best practices for Amazon S3 destinations

Deleting an S3 bucket that's configured as a destination without removing the destination from your function's configuration can create a security risk. If another user knows your destination bucket's name, they can recreate the bucket in their AWS account. Records of failed invocations will be sent to their bucket, potentially exposing data from your function.

The following example shows an IAM policy that limits your function's s3:PutObject permissions to buckets in your account. This policy also gives Lambda the s3:ListBucket permission it needs to use an S3 bucket as a destination.

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Sid": "S3BucketResourceAccountWrite",
            "Effect": "Allow",
            "Action": [
                "s3:PutObject",
                "s3:ListBucket"
            ],
            "Resource": "arn:aws:s3:::*/*",
            "Condition": {
                "StringEquals": {
                    "s3:ResourceAccount": "111122223333"
                }
            }
        }
    ]
}

To add a permissions policy to your funcion's execution role using the AWS Management Console or AWS CLI, refer to the instructions in the following procedures:

Example invocation record

When an invocation matches the condition, Lambda sends a JSON document with details about the invocation to the destination. The following example shows an invocation record for an event that failed three processing attempts due to a function error.

{
    "version": "1.0",
    "timestamp": "2019-11-14T18:16:05.568Z",
    "requestContext": {
        "requestId": "e4b46cbf-b738-xmpl-8880-a18cdf61200e",
        "functionArn": "arn:aws:lambda:us-east-1:123456789012:function:my-function:$LATEST",
        "condition": "RetriesExhausted",
        "approximateInvokeCount": 3
    },
    "requestPayload": {
        "ORDER_IDS": [
            "9e07af03-ce31-4ff3-xmpl-36dce652cb4f",
            "637de236-e7b2-464e-xmpl-baf57f86bb53",
            "a81ddca6-2c35-45c7-xmpl-c3a03a31ed15"
        ]
    },
    "responseContext": {
        "statusCode": 200,
        "executedVersion": "$LATEST",
        "functionError": "Unhandled"
    },
    "responsePayload": {
        "errorMessage": "RequestId: e4b46cbf-b738-xmpl-8880-a18cdf61200e Process exited before completing request"
    }
}

The invocation record contains details about the event, the response, and the reason that the record was sent.

Tracing requests to destinations

You can use AWS X-Ray to see a connected view of each request as it's queued, processed by a Lambda function, and passed to the destination service. When you activate X-Ray tracing for a function or a service that invokes a function, Lambda adds an X-Ray header to the request and passes the header to the destination service. Traces from upstream services are automatically linked to traces from downstream Lambda functions and destination services, creating an end-to-end view of the entire application. For more information about tracing, see Visualize Lambda function invocations using AWS X-Ray.

Adding a dead-letter queue

As an alternative to an on-failure destination, you can configure your function with a dead-letter queue to save discarded events for further processing. A dead-letter queue acts the same as an on-failure destination in that it is used when an event fails all processing attempts or expires without being processed. However, you can only add or remove a dead-letter queue at the function level. Function versions use the same dead-letter queue settings as the unpublished version ($LATEST). On-failure destinations also support additional targets and include details about the function's response in the invocation record.

To reprocess events in a dead-letter queue, you can set it as an event source for your Lambda function. Alternatively, you can manually retrieve the events.

You can choose an Amazon SQS standard queue or Amazon SNS standard topic for your dead-letter queue. FIFO queues and Amazon SNS FIFO topics are not supported.

To send events to a queue or topic, your function needs additional permissions. Add a policy with the required permissions to your function's execution role.

If the target queue or topic is encrypted with a customer managed key, the execution role must also be a user in the key's resource-based policy.

After creating the target and updating your function's execution role, add the dead-letter queue to your function. You can configure multiple functions to send events to the same target.

Lambda sends the event to the dead-letter queue as-is, with additional information in attributes. You can use this information to identify the error that the function returned, or to correlate the event with logs or an AWS X-Ray trace.

If Lambda can't send a message to the dead-letter queue, it deletes the event and emits the DeadLetterErrors metric. This can happen because of lack of permissions, or if the total size of the message exceeds the limit for the target queue or topic. For example, say that an Amazon SNS notification with a body close to 256 KB in size triggers a function that results in an error. In that case, the event data that Amazon SNS adds, combined with the attributes that Lambda adds, can cause the message to exceed the maximum size allowed in the dead-letter queue.

If you're using Amazon SQS as an event source, configure a dead-letter queue on the Amazon SQS queue itself and not on the Lambda function. For more information, see Using Lambda with Amazon SQS.