Using Lambda with Amazon MSK

Amazon Managed Streaming for Apache Kafka (Amazon MSK) is a fully managed service that you can use to build and run applications that use Apache Kafka to process streaming data. Amazon MSK simplifies the setup, scaling, and management of clusters running Kafka. Amazon MSK also makes it easier to configure your application for multiple Availability Zones and for security with Amazon Identity and Access Management (IAM). Amazon MSK supports multiple open-source versions of Kafka.

Amazon MSK as an event source operates similarly to using Amazon Simple Queue Service (Amazon SQS) or Amazon Kinesis. Lambda internally polls for new messages from the event source and then synchronously invokes the target Lambda function. Lambda reads the messages in batches and provides these to your function as an event payload. The maximum batch size is configurable (the default is 100 messages). For more information, see Batching behavior.

Lambda reads the messages sequentially for each partition. A single Lambda payload can contain messages from multiple partitions. After Lambda processes each batch, it commits the offsets of the messages in that batch. If your function returns an error for any of the messages in a batch, Lambda retries the whole batch of messages until processing succeeds or the messages expire.

For an example of how to configure Amazon MSK as an event source, see Using Amazon MSK as an event source for Amazon Lambda on the Amazon Compute Blog. For a complete tutorial, see Amazon MSK Lambda Integration in the Amazon MSK Labs.

Example event

Lambda sends the batch of messages in the event parameter when it invokes your function. The event payload contains an array of messages. Each array item contains details of the Amazon MSK topic and partition identifier, together with a timestamp and a base64-encoded message.

{
   "eventSource":"aws:kafka",
   "eventSourceArn":"arn:aws:kafka:sa-east-1:123456789012:cluster/vpc-2priv-2pub/751d2973-a626-431c-9d4e-d7975eb44dd7-2",
   "bootstrapServers":"b-2.demo-cluster-1.a1bcde.c1.kafka.us-east-1.amazonaws.com:9092,b-1.demo-cluster-1.a1bcde.c1.kafka.us-east-1.amazonaws.com:9092",
   "records":{
      "mytopic-0":[
         {
            "topic":"mytopic",
            "partition":0,
            "offset":15,
            "timestamp":1545084650987,
            "timestampType":"CREATE_TIME",
            "key":"abcDEFghiJKLmnoPQRstuVWXyz1234==",
            "value":"SGVsbG8sIHRoaXMgaXMgYSB0ZXN0Lg==",
            "headers":[
               {
                  "headerKey":[
                     104,
                     101,
                     97,
                     100,
                     101,
                     114,
                     86,
                     97,
                     108,
                     117,
                     101
                  ]
               }
            ]
         }
      ]
   }
}

MSK cluster authentication

Lambda needs permission to access the Amazon MSK cluster, retrieve records, and perform other tasks. Amazon MSK supports several options for controlling client access to the MSK cluster.

Unauthenticated access

If no clients access the cluster over the internet, you can use unauthenticated access.

SASL/SCRAM authentication

Amazon MSK supports Simple Authentication and Security Layer/Salted Challenge Response Authentication Mechanism (SASL/SCRAM) authentication with Transport Layer Security (TLS) encryption. For Lambda to connect to the cluster, you store the authentication credentials (user name and password) in an Amazon Secrets Manager secret.

For more information about using Secrets Manager, see User name and password authentication with Amazon Secrets Manager in the Amazon Managed Streaming for Apache Kafka Developer Guide.

Amazon MSK doesn't support SASL/PLAIN authentication.

IAM role-based authentication

You can use IAM to authenticate the identity of clients that connect to the MSK cluster. If IAM auth is active on your MSK cluster, and you don't provide a secret for auth, Lambda automatically defaults to using IAM auth. To create and deploy user or role-based policies, use the IAM console or API. For more information, see IAM access control in the Amazon Managed Streaming for Apache Kafka Developer Guide.

To allow Lambda to connect to the MSK cluster, read records, and perform other required actions, add the following permissions to your function's execution role.

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "kafka-cluster:Connect",
                "kafka-cluster:DescribeGroup",
                "kafka-cluster:AlterGroup",
                "kafka-cluster:DescribeTopic",
                "kafka-cluster:ReadData",
                "kafka-cluster:DescribeClusterDynamicConfiguration"
            ],
            "Resource": [
                "arn:aws:kafka:region:account-id:cluster/cluster-name/cluster-uuid",
                "arn:aws:kafka:region:account-id:topic/cluster-name/cluster-uuid/topic-name",
                "arn:aws:kafka:region:account-id:group/cluster-name/cluster-uuid/consumer-group-id"
            ]
        }
    ]
}       

You can scope these permissions to a specific cluster, topic, and group. For more information, see the Amazon MSK Kafka actions in the Amazon Managed Streaming for Apache Kafka Developer Guide.

Mutual TLS authentication

Mutual TLS (mTLS) provides two-way authentication between the client and server. The client sends a certificate to the server for the server to verify the client, and the server sends a certificate to the client for the client to verify the server.

For Amazon MSK, Lambda acts as the client. You configure a client certificate (as a secret in Secrets Manager) to authenticate Lambda with the brokers in your MSK cluster. The client certificate must be signed by a CA in the server's trust store. The MSK cluster sends a server certificate to Lambda to authenticate the brokers with Lambda. The server certificate must be signed by a certificate authority (CA) that's in the Amazon trust store.

For instructions on how to generate a client certificate, see Introducing mutual TLS authentication for Amazon MSK as an event source.

Amazon MSK doesn't support self-signed server certificates, because all brokers in Amazon MSK use public certificates signed by Amazon Trust Services CAs, which Lambda trusts by default.

For more information about mTLS for Amazon MSK, see Mutual TLS Authentication in the Amazon Managed Streaming for Apache Kafka Developer Guide.

Configuring the mTLS secret

The CLIENT_CERTIFICATE_TLS_AUTH secret requires a certificate field and a private key field. For an encrypted private key, the secret requires a private key password. Both the certificate and private key must be in PEM format.

The certificate field must contain a list of certificates, beginning with the client certificate, followed by any intermediate certificates, and ending with the root certificate. Each certificate must start on a new line with the following structure:

-----BEGIN CERTIFICATE-----  
        <certificate contents>
-----END CERTIFICATE-----      

Secrets Manager supports secrets up to 65,536 bytes, which is enough space for long certificate chains.

The private key must be in PKCS #8 format, with the following structure:

-----BEGIN PRIVATE KEY-----  
         <private key contents>
-----END PRIVATE KEY-----            

For an encrypted private key, use the following structure:

-----BEGIN ENCRYPTED PRIVATE KEY-----  
          <private key contents>
-----END ENCRYPTED PRIVATE KEY-----           

The following example shows the contents of a secret for mTLS authentication using an encrypted private key. For an encrypted private key, you include the private key password in the secret.

{
 "privateKeyPassword": "testpassword",
 "certificate": "-----BEGIN CERTIFICATE-----
MIIE5DCCAsygAwIBAgIRAPJdwaFaNRrytHBto0j5BA0wDQYJKoZIhvcNAQELBQAw
...
j0Lh4/+1HfgyE2KlmII36dg4IMzNjAFEBZiCRoPimO40s1cRqtFHXoal0QQbIlxk
cmUuiAii9R0=
-----END CERTIFICATE-----
-----BEGIN CERTIFICATE-----
MIIFgjCCA2qgAwIBAgIQdjNZd6uFf9hbNC5RdfmHrzANBgkqhkiG9w0BAQsFADBb
...
rQoiowbbk5wXCheYSANQIfTZ6weQTgiCHCCbuuMKNVS95FkXm0vqVD/YpXKwA/no
c8PH3PSoAaRwMMgOSA2ALJvbRz8mpg==
-----END CERTIFICATE-----",
 "privateKey": "-----BEGIN ENCRYPTED PRIVATE KEY-----
MIIFKzBVBgkqhkiG9w0BBQ0wSDAnBgkqhkiG9w0BBQwwGgQUiAFcK5hT/X7Kjmgp
...
QrSekqF+kWzmB6nAfSzgO9IaoAaytLvNgGTckWeUkWn/V0Ck+LdGUXzAC4RxZnoQ
zp2mwJn2NYB7AZ7+imp0azDZb+8YG2aUCiyqb6PnnA==
-----END ENCRYPTED PRIVATE KEY-----"
} 

How Lambda chooses a bootstrap broker

Lambda chooses a bootstrap broker based on the authentication methods available on your cluster, and whether you provide a secret for authentication. If you provide a secret for mTLS or SASL/SCRAM, Lambda automatically chooses that auth method. If you don't provide a secret, Lambda selects the strongest auth method that's active on your cluster. The following is the order of priority in which Lambda selects a broker, from strongest to weakest auth:

Managing API access and permissions

In addition to accessing the Amazon MSK cluster, your function needs permissions to perform various Amazon MSK API actions. You add these permissions to the function's execution role. If your users need access to any of the Amazon MSK API actions, add the required permissions to the identity policy for the user or role.

You can add each of the following permissions to your execution role manually. Alternatively, you can attach the Amazon managed policy AWSLambdaMSKExecutionRole to your execution role. The AWSLambdaMSKExecutionRole policy contains all required API actions and VPC permissions listed below.

Required Lambda function execution role permissions

To create and store logs in a log group in Amazon CloudWatch Logs, your Lambda function must have the following permissions in its execution role:

For Lambda to access your Amazon MSK cluster on your behalf, your Lambda function must have the following permissions in its execution role:

You only need to add one of either kafka:DescribeCluster or kafka:DescribeClusterV2. For provisioned MSK clusters, either permission works. For serverless MSK clusters, you must use kafka:DescribeClusterV2.

VPC permissions

If only users within a VPC can access your Amazon MSK cluster, your Lambda function must have permission to access your Amazon VPC resources. These resources include your VPC, subnets, security groups, and network interfaces. To access these resources, your function's execution role must have the following permissions. These permissions are included in the AWSLambdaMSKExecutionRole Amazon managed policy.

Optional Lambda function permissions

Your Lambda function might also need permissions to:

Secrets Manager and Amazon KMS permissions

Depending on the type of access control that you're configuring for your Amazon MSK brokers, your Lambda function might need permission to access your SCRAM secret (if using SASL/SCRAM authentication), or Secrets Manager secret to decrypt your Amazon KMS customer managed key. To access these resources, your function's execution role must have the following permissions:

Adding permissions to your execution role

Follow these steps to add the Amazon managed policy AWSLambdaMSKExecutionRole to your execution role using the IAM console.

Granting users access with an IAM policy

By default, users and roles don't have permission to perform Amazon MSK API operations. To grant access to users in your organization or account, you can add or update an identity-based policy. For more information, see Amazon MSK Identity-Based Policy Examples in the Amazon Managed Streaming for Apache Kafka Developer Guide.

Authentication and authorization errors

If any of the permissions required to consume data from the Amazon MSK cluster are missing, Lambda displays one of the following error messages in the event source mapping under LastProcessingResult.

Cluster failed to authorize Lambda

For SASL/SCRAM or mTLS, this error indicates that the provided user doesn't have all of the following required Kafka access control list (ACL) permissions:

For IAM access control, your function's execution role is missing one or more of the permissions required to access the group or topic. Review the list of required permissions in IAM role-based authentication.

When you create either Kafka ACLs or an IAM policy with the required Kafka cluster permissions, specify the topic and group as resources. The topic name must match the topic in the event source mapping. The group name must match the event source mapping's UUID.

After you add the required permissions to the execution role, it might take several minutes for the changes to take effect.

SASL authentication failed

For SASL/SCRAM, this error indicates that the provided user name and password aren't valid.

For IAM access control, the execution role is missing the kafka-cluster:Connect permission for the MSK cluster. Add this permission to the role and specify the cluster's Amazon Resource Name (ARN) as a resource.

You might see this error occurring intermittently. The cluster rejects connections after the number of TCP connections exceeds the Amazon MSK service quota. Lambda backs off and retries until a connection is successful. After Lambda connects to the cluster and polls for records, the last processing result changes to OK.

Server failed to authenticate Lambda

This error indicates that the Amazon MSK Kafka brokers failed to authenticate with Lambda. This can occur for any of the following reasons:

Provided certificate or private key is invalid

This error indicates that the Amazon MSK consumer couldn't use the provided certificate or private key. Make sure that the certificate and key use PEM format, and that the private key encryption uses a PBES1 algorithm.

Network configuration

For Lambda to use your Kafka cluster as an event source, it needs access to the Amazon VPC your cluster resides in. We recommend that you deploy Amazon PrivateLink VPC endpoints for Lambda to access your VPC. Deploy endpoints for Lambda and Amazon Security Token Service (Amazon STS). If the broker uses authentication, also deploy a VPC endpoint for Secrets Manager. If you configured an on-failure destination, also deploy a VPC endpoint for the destination service.

Alternatively, ensure that the VPC associated with your Kafka cluster includes one NAT gateway per public subnet. For more information, see Enable internet access for VPC-connected Lambda functions.

If you use VPC endpoints, you must also configure them to enable private DNS names.

When you create an event source mapping for an MSK cluster, Lambda checks whether Elastic Network Interfaces (ENIs) are already present for the subnets and security groups of your cluster’s VPC. If Lambda finds existing ENIs, it attempts to re-use them. Otherwise, Lambda creates new ENIs to connect to the event source and invoke your function.

Your Amazon VPC configuration is discoverable through the Amazon MSK API. You don't need to configure it during setup using the create-event-source-mapping command.

For more information about configuring the network, see Setting up Amazon Lambda with an Apache Kafka cluster within a VPC on the Amazon Compute Blog.

VPC security group rules

Configure the security groups for the Amazon VPC containing your cluster with the following rules (at minimum):

Working with VPC endpoints

When you use VPC endpoints, API calls to invoke your function are routed through these endpoints using the ENIs. The Lambda service principal needs to call sts:AssumeRole and lambda:InvokeFunction on any roles and functions that use those ENIs.

By default, VPC endpoints have IAM policies which are open. Best practice is to restrict these policies to allow only specific principals to perform the needed actions using that endpoint. To ensure that your event source mapping is able to invoke your Lambda function, the VPC endpoint policy must allow the Lambda service principle to call sts:AssumeRole and lambda:InvokeFunction. Restricting your VPC endpoint policies to allow only API calls originating within your organization prevents the event source mapping from functioning properly.

The following example VPC endpoint policies show how to grant the required access to the Lambda service principal for the Amazon STS and Lambda endpoints.

{
    "Statement": [
        {
            "Action": "sts:AssumeRole",
            "Effect": "Allow",
            "Principal": {
                "Service": [
                    "lambda.amazonaws.com"
                ]
            },
            "Resource": "*"
        }
    ]
}

{
    "Statement": [
        {
            "Action": "lambda:InvokeFunction",
            "Effect": "Allow",
            "Principal": {
                "Service": [
                    "lambda.amazonaws.com"
                ]
            },
            "Resource": "*"
        }
    ]
}

If your Kafka broker uses authentication, you can also restrict the VPC endpoint policy for the Secrets Manager endpoint. To call the Secrets Manager API, Lambda uses your function role, not the Lambda service principal. The following example shows a Secrets Manager endpoint policy.

{
    "Statement": [
        {
            "Action": "secretsmanager:GetSecretValue",
            "Effect": "Allow",
            "Principal": {
                "AWS": [
                    "customer_function_execution_role_arn"
                ]
            },
            "Resource": "customer_secret_arn"
        }
    ]
}

If you have an on-failure destination configured, Lambda also uses your function’s role to call either s3:PutObject, sns:Publish, or sqs:sendMessage using the Lambda-managed ENIs.

Adding Amazon MSK as an event source

To create an event source mapping, add Amazon MSK as a Lambda function trigger using the Lambda console, an Amazon SDK, or the Amazon Command Line Interface (Amazon CLI). Note that when you add Amazon MSK as a trigger, Lambda assumes the VPC settings of the Amazon MSK cluster, not the Lambda function's VPC settings.

This section describes how to create an event source mapping using the Lambda console and the Amazon CLI.

Prerequisites

Customizable consumer group ID

When setting up Kafka as an event source, you can specify a consumer group ID. This consumer group ID is an existing identifier for the Kafka consumer group that you want your Lambda function to join. You can use this feature to seamlessly migrate any ongoing Kafka record processing setups from other consumers to Lambda.

If you specify a consumer group ID and there are other active pollers within that consumer group, Kafka distributes messages across all consumers. In other words, Lambda doesn't receive all message for the Kafka topic. If you want Lambda to handle all messages in the topic, turn off any other pollers in that consumer group.

Additionally, if you specify a consumer group ID, and Kafka finds a valid existing consumer group with the same ID, Lambda ignores the StartingPosition parameter for your event source mapping. Instead, Lambda begins processing records according to the committed offset of the consumer group. If you specify a consumer group ID, and Kafka cannot find an existing consumer group, then Lambda configures your event source with the specified StartingPosition.

The consumer group ID that you specify must be unique among all your Kafka event sources. After creating a Kafka event source mapping with the consumer group ID specified, you cannot update this value.

Adding an Amazon MSK trigger (console)

Follow these steps to add your Amazon MSK cluster and a Kafka topic as a trigger for your Lambda function.

Adding an Amazon MSK trigger (Amazon CLI)

Use the following example Amazon CLI commands to create and view an Amazon MSK trigger for your Lambda function.

Creating a trigger using the Amazon CLI

aws lambda create-event-source-mapping \
  --event-source-arn arn:aws-cn:kafka:us-east-1:111122223333:cluster/my-cluster/fc2f5bdf-fd1b-45ad-85dd-15b4a5a6247e-2 \
  --topics AWSKafkaTopic \
  --starting-position LATEST \
  --function-name my-kafka-function

aws lambda create-event-source-mapping \
  --event-source-arn arn:aws-cn:kafka:us-east-1:111122223333:cluster/my-cluster/fc2f5bdf-fd1b-45ad-85dd-15b4a5a6247e-2 \
  --topics AWSKafkaTopic \
  --starting-position LATEST \
  --function-name my-kafka-function
  --source-access-configurations '[{"Type": "SASL_SCRAM_512_AUTH","URI": "arn:aws:secretsmanager:us-east-1:111122223333:secret:my-secret"}]'

aws lambda create-event-source-mapping \
  --event-source-arn arn:aws-cn:kafka:us-east-1:111122223333:cluster/my-cluster/fc2f5bdf-fd1b-45ad-85dd-15b4a5a6247e-2 \
  --topics AWSKafkaTopic \
  --starting-position LATEST \
  --function-name my-kafka-function
  --source-access-configurations '[{"Type": "CLIENT_CERTIFICATE_TLS_AUTH","URI": "arn:aws:secretsmanager:us-east-1:111122223333:secret:my-secret"}]'

For more information, see the CreateEventSourceMapping API reference documentation.

Viewing the status using the Amazon CLI

The following example uses the get-event-source-mapping Amazon CLI command to describe the status of the event source mapping that you created.

aws lambda get-event-source-mapping \
  --uuid 6d9bce8e-836b-442c-8070-74e77903c815

Creating cross-account event source mappings

You can use multi-VPC private connectivity to connect a Lambda function to a provisioned MSK cluster in a different Amazon Web Services account. Multi-VPC connectivity uses Amazon PrivateLink, which keeps all traffic within the Amazon network.

To create a cross-account event source mapping, you must first configure multi-VPC connectivity for the MSK cluster. When you create the event source mapping, use the managed VPC connection ARN instead of the cluster ARN, as shown in the following examples. The CreateEventSourceMapping operation also differs depending on which authentication type the MSK cluster uses.

aws lambda create-event-source-mapping \
  --event-source-arn arn:aws:kafka:us-east-1:111122223333:vpc-connection/444455556666/my-cluster-name/51jn98b4-0a61-46cc-b0a6-61g9a3d797d5-7 \
  --topics AWSKafkaTopic \
  --starting-position LATEST \
  --function-name my-kafka-function

aws lambda create-event-source-mapping \
  --event-source-arn arn:aws:kafka:us-east-1:111122223333:vpc-connection/444455556666/my-cluster-name/51jn98b4-0a61-46cc-b0a6-61g9a3d797d5-7 \
  --topics AWSKafkaTopic \
  --starting-position LATEST \
  --function-name my-kafka-function \
  --source-access-configurations '[{"Type": "SASL_SCRAM_512_AUTH","URI": "arn:aws:secretsmanager:us-east-1:444455556666:secret:my-secret"}]'

aws lambda create-event-source-mapping \
  --event-source-arn arn:aws:kafka:us-east-1:111122223333:vpc-connection/444455556666/my-cluster-name/51jn98b4-0a61-46cc-b0a6-61g9a3d797d5-7 \
  --topics AWSKafkaTopic \
  --starting-position LATEST \
  --function-name my-kafka-function \
  --source-access-configurations '[{"Type": "CLIENT_CERTIFICATE_TLS_AUTH","URI": "arn:aws:secretsmanager:us-east-1:444455556666:secret:my-secret"}]'

On-failure destinations

To retain records of failed event source mapping invocations, add a destination to your function's event source mapping. Each record sent to the destination is a JSON document with metadata about the failed invocation. You can configure any Amazon SNS topic, Amazon SQS queue, or S3 bucket as a destination. Your execution role must have permissions for the destination:

Additionally, if you configured a KMS key on your destination, Lambda needs the following permissions depending on the destination type:

To configure an on-failure destination using the console, follow these steps:

You can also configure an on-failure destination using the Amazon CLI. For example, the following create-event-source-mapping command adds an event source mapping with an SQS on-failure destination to MyFunction:

aws lambda create-event-source-mapping \
--function-name "MyFunction" \
--event-source-arn arn:aws-cn:kafka:us-east-1:123456789012:cluster/vpc-2priv-2pub/751d2973-a626-431c-9d4e-d7975eb44dd7-2 \
--destination-config '{"OnFailure": {"Destination": "arn:aws:sqs:us-east-1:123456789012:dest-queue"}}'

The following update-event-source-mapping command adds an S3 on-failure destination to the event source associated with the input uuid:

aws lambda update-event-source-mapping \
--uuid f89f8514-cdd9-4602-9e1f-01a5b77d449b \
--destination-config '{"OnFailure": {"Destination": "arn:aws:s3:::dest-bucket"}}'

To remove a destination, supply an empty string as the argument to the destination-config parameter:

aws lambda update-event-source-mapping \
--uuid f89f8514-cdd9-4602-9e1f-01a5b77d449b \
--destination-config '{"OnFailure": {"Destination": ""}}'

SNS and SQS example invocation record

The following example shows what Lambda sends to an SNS topic or SQS queue destination for a failed Kafka event source invocation. Each of the keys under recordsInfo contains both the Kafka topic and partition, separated by a hyphen. For example, for the key "Topic-0", Topic is the Kafka topic, and 0 is the partition. For each topic and partition, you can use the offsets and timestamp data to find the original invocation records.

{
    "requestContext": {
        "requestId": "316aa6d0-8154-xmpl-9af7-85d5f4a6bc81",
        "functionArn": "arn:aws:lambda:us-east-1:123456789012:function:myfunction",
        "condition": "RetryAttemptsExhausted" | "MaximumPayloadSizeExceeded",
        "approximateInvokeCount": 1
    },
    "responseContext": { // null if record is MaximumPayloadSizeExceeded
        "statusCode": 200,
        "executedVersion": "$LATEST",
        "functionError": "Unhandled"
    },
    "version": "1.0",
    "timestamp": "2019-11-14T00:38:06.021Z",
    "KafkaBatchInfo": {
        "batchSize": 500,
        "eventSourceArn": "arn:aws-cn:kafka:us-east-1:123456789012:cluster/vpc-2priv-2pub/751d2973-a626-431c-9d4e-d7975eb44dd7-2",
        "bootstrapServers": "...",
        "payloadSize": 2039086, // In bytes
        "recordsInfo": {
            "Topic-0": {
                "firstRecordOffset": "49601189658422359378836298521827638475320189012309704722",
                "lastRecordOffset": "49601189658422359378836298522902373528957594348623495186",
                "firstRecordTimestamp": "2019-11-14T00:38:04.835Z",
                "lastRecordTimestamp": "2019-11-14T00:38:05.580Z",
            },
            "Topic-1": {
                "firstRecordOffset": "49601189658422359378836298521827638475320189012309704722",
                "lastRecordOffset": "49601189658422359378836298522902373528957594348623495186",
                "firstRecordTimestamp": "2019-11-14T00:38:04.835Z",
                "lastRecordTimestamp": "2019-11-14T00:38:05.580Z",
            }
        }
    }
}

S3 destination example invocation record

For S3 destinations, Lambda sends the entire invocation record along with the metadata to the destination. The following example shows that Lambda sends to an S3 bucket destination for a failed Kafka event source invocation. In addition to all of the fields from the previous example for SQS and SNS destinations, the payload field contains the original invocation record as an escaped JSON string.

{
    "requestContext": {
        "requestId": "316aa6d0-8154-xmpl-9af7-85d5f4a6bc81",
        "functionArn": "arn:aws:lambda:us-east-1:123456789012:function:myfunction",
        "condition": "RetryAttemptsExhausted" | "MaximumPayloadSizeExceeded",
        "approximateInvokeCount": 1
    },
    "responseContext": { // null if record is MaximumPayloadSizeExceeded
        "statusCode": 200,
        "executedVersion": "$LATEST",
        "functionError": "Unhandled"
    },
    "version": "1.0",
    "timestamp": "2019-11-14T00:38:06.021Z",
    "KafkaBatchInfo": {
        "batchSize": 500,
        "eventSourceArn": "arn:aws-cn:kafka:us-east-1:123456789012:cluster/vpc-2priv-2pub/751d2973-a626-431c-9d4e-d7975eb44dd7-2",
        "bootstrapServers": "...",
        "payloadSize": 2039086, // In bytes
        "recordsInfo": {
            "Topic-0": {
                "firstRecordOffset": "49601189658422359378836298521827638475320189012309704722",
                "lastRecordOffset": "49601189658422359378836298522902373528957594348623495186",
                "firstRecordTimestamp": "2019-11-14T00:38:04.835Z",
                "lastRecordTimestamp": "2019-11-14T00:38:05.580Z",
            },
            "Topic-1": {
                "firstRecordOffset": "49601189658422359378836298521827638475320189012309704722",
                "lastRecordOffset": "49601189658422359378836298522902373528957594348623495186",
                "firstRecordTimestamp": "2019-11-14T00:38:04.835Z",
                "lastRecordTimestamp": "2019-11-14T00:38:05.580Z",
            }
        }
    },
    "payload": "<Whole Event>" // Only available in S3
}

Auto scaling of the Amazon MSK event source

When you initially create an Amazon MSK event source, Lambda allocates one consumer to process all partitions in the Kafka topic. Each consumer has multiple processors running in parallel to handle increased workloads. Additionally, Lambda automatically scales up or down the number of consumers, based on workload. To preserve message ordering in each partition, the maximum number of consumers is one consumer per partition in the topic.

In one-minute intervals, Lambda evaluates the consumer offset lag of all the partitions in the topic. If the lag is too high, the partition is receiving messages faster than Lambda can process them. If necessary, Lambda adds or removes consumers from the topic. The scaling process of adding or removing consumers occurs within three minutes of evaluation.

If your target Lambda function is throttled, Lambda reduces the number of consumers. This action reduces the workload on the function by reducing the number of messages that consumers can retrieve and send to the function.

To monitor the throughput of your Kafka topic, view the Offset lag metric Lambda emits while your function processes records.

To check how many function invocations occur in parallel, you can also monitor the concurrency metrics for your function.

Polling and stream starting positions

Be aware that stream polling during event source mapping creation and updates is eventually consistent.

This behavior means that if you specify LATEST as the starting position for the stream, the event source mapping could miss events during creation or updates. To ensure that no events are missed, specify the stream starting position as TRIM_HORIZON or AT_TIMESTAMP.

Amazon CloudWatch metrics

Lambda emits the OffsetLag metric while your function processes records. The value of this metric is the difference in offset between the last record written to the Kafka event source topic and the last record that your function's consumer group processed. You can use OffsetLag to estimate the latency between when a record is added and when your consumer group processes it.

An increasing trend in OffsetLag can indicate issues with pollers in your function's consumer group. For more information, see Working with Lambda function metrics.

Amazon MSK configuration parameters

All Lambda event source types share the same CreateEventSourceMapping and UpdateEventSourceMapping API operations. However, only some of the parameters apply to Amazon MSK.