Amazon MSK Mutual TLS authentication

Estimated time to complete 20-30 minutes.

Overview

The Zilla Plus for Amazon MSKopen in new window Secure Public Access proxy lets authorized Kafka clients connect, publish messages and subscribe to topics in your Amazon MSK cluster via the internet.

In this guide we will deploy the Zilla Plus for Amazon MSK Secure Public Access proxy and showcase globally trusted public internet connectivity to an MSK cluster from a Kafka client, using the custom wildcard domain *.example.aklivity.io. Kafka clients will use TLS client certificates to verify trusted client identity.

AWS services used

Service	Required	Usage	Quota
Resource Groups and Tagging	Yes	Startup only	Noneopen in new window
Secrets Manager	Yes	Startup only	Not reachedopen in new window
Certificate Manager	No Private key and certificate can be inline in Secrets Manager instead	Startup only	Not reachedopen in new window
Private Certificate Manager	No Private key and certificate can be inline in Secrets Manager instead	Startup only	Not reachedopen in new window

Default AWS Service Quotasopen in new window are recommended.

Prerequisites

Before setting up internet access to your MSK Cluster, you will need the following:

an MSK Cluster configured for TLS encrypted client access and TLS client authentication
an VPC security group for Zilla proxies
an IAM security role for Zilla proxies
subscription to Zilla Plus for Amazon MSK via AWS Marketplace
permission to modify global DNS records for a custom domain
permission to generate client certificates signed by a private certificate authority

Tips

Check out the Troubleshooting guide if you run into any issues.

Create Certificate Authority for mTLS

This creates a new private certificate authority in ACM.

Follow the Create Certificate Authority to create a private certificate authority to verify TLS client authentication.

Distinguished Name
Common Name (CN)
```
Mutual Authentication CA
```

Create the MSK Cluster

This creates your MSK cluster in preparation for secure access via the internet.

An MSK cluster is needed for secure remote access via the internet. You can skip this step if you have already created an MSK cluster with equivalent configuration.

Follow the Create MSK Cluster guide to setup the a new MSK cluster. We will use the below resource names to reference the AWS resources needed in this guide.

Cluster Name: my-msk-cluster
Access control methods: TLS client certificates
AWS Private CAs: Mutual Authentication CA
VPC: my-msk-cluster-vpc
Subnet: my-msk-cluster-subnet-*
Route tables: my-msk-cluster-rtb-*
Internet gateway: my-msk-cluster-igw

Create Client Certificate for mTLS

This allows an authorized Kafka client to connect directly to your MSK cluster with Mutual TLS (mTLS). Follow the Create Client Certificate to create a private certificate authority.

You can create additional client certificates for each different authorized client identity that will connect via the internet to your Zilla proxy deployment.

Update the security settingsopen in new window of your MSK cluster. Update your Access control methods to include

Common Name: client-1
Private Certificate Authority: Mutual Authentication CA

Create the Zilla proxy security group

This creates your Zilla proxy security group to allow Kafka clients and SSH access.

A VPC security group is needed for the Zilla proxies when they are launched.

Follow the Create Security Groupopen in new window wizard with the following parameters and defaults. This creates your Zilla proxy security group to allow Kafka clients and SSH access.

VPC: my-msk-cluster-vpc
Name: my-zilla-proxy-sg
Description: Kafka clients and SSH access
Add Inbound Rule
- Type: CUSTOM TCP
- Port Range: 9094
- Source type: Anywhere-IPv4
Add Inbound Rule
- Type: SSH
- Source type: My IP
Add Outbound Rule (if not exists)
- Type: All traffic
- Destination: Anywhere-IPv4
Create the Security Group

Check your network settings

Your IP may be different when you SSH into the EC2 instance. VPNs and other networking infrastructure may cause the My IP inbound rule to fail. Instead, you can use one of the other ways AWS provides to execute commands in an EC2 instance.

Update the default security group rules

This allows the Zilla proxies to communicate with your MSK cluster.

Navigate to the VPC Management Console Security Groupsopen in new window table.

Check your selected region

Make sure you have selected the desired region, ex: US East (N. Virginia) us-east-1.

Filter the security groups by selecting a VPC and select the default security group.

VPC: my-msk-cluster-vpc
Security Group: default

Add a Custom TCP Rule

Add this Inbound Rule to allow the Zilla proxies to communicate with the MSK cluster.

Type: Custom TCP
Port Range: 9094
Source type: Custom
Source: my-zilla-proxy-sg

Create the Zilla proxy IAM security role

This creates an IAM security role to enable the required AWS services for the Zilla proxies.

Follow the Create IAM Role guide to create an IAM security role with the following parameters:

Name

aklivity-zilla-proxy

Policies

AWSCertificateManagerReadOnly
AWSCertificateManagerPrivateCAReadOnly
ResourceGroupsandTagEditorReadOnlyAccess

IAM role Inline Policies

This creates an IAM security role to enable the required AWS services for the Zilla proxies.

Name

MSKProxySecretsManagerRead

JSON Summary

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Sid": "VisualEditor0",
      "Effect": "Allow",
      "Action": [
        "secretsmanager:GetSecretValue",
        "secretsmanager:DescribeSecret"
      ],
      "Resource": [
        "arn:aws:secretsmanager:*:*:secret:wildcard.example.aklivity.io-*",
        "arn:aws:secretsmanager:*:*:secret:client-*"
      ]
    }
  ]
}

Note

This example pattern requires all trusted client certificate key secrets to be named client-*.

If you used a different secret name for your certificate key.

Replace wildcard.example.aklivity.io in the resource regular expression for:

MSKProxySecretsManagerRead

The Zilla Plus for Amazon MSKopen in new window is available through the AWS Marketplace. You can skip this step if you have already subscribed to Zilla Plus for Amazon MSK via AWS Marketplace.

To get started, visit the Proxy's Marketplace Product Pageopen in new window and Subscribe to the offering. You should now see Zilla Plus for Amazon MSK listed in your AWS Marketplaceopen in new window subscriptions.

Create the Public TLS Server Certificate

We need a Public TLS Server Certificate for your custom DNS wildcard domain that can be trusted by a Kafka Client from anywhere.

Follow the Create Server Certificate (LetsEncrypt) guide to create a new TLS Server Certificate. Use your own custom wildcard DNS domain in place of the example wildcard domain *.example.aklivity.io.

Info

Note the server certificate secret ARN as we will need to reference it from the Secure Public Access CloudFormation template.

Deploy the Zilla Plus Secure Public Access Proxy

This initiates deployment of the Zilla Plus for Amazon MSK stack via CloudFormation.

Navigate to your AWS Marketplaceopen in new window subscriptions and select Zilla Plus for Amazon MSK to show the manage subscription page.

From the Agreement section > Actions menu > select Launch CloudFormation stack
Select the Secure Public Access (mTLS) fulfillment option
Make sure you have selected the desired region selected, such as us-east-1
Click Continue to Launch
- Choose the action Launch CloudFormation

Click Launch to complete the Create stack wizard with the following details:

Step 1. Create Stack

Prepare template: Template is ready
Specify template: Amazon S3 URL
- Amazon S3 URL: (auto-filled)

Step 2. Specify stack details

Stack name

my-zilla-proxy

Parameters:

Network Configuration
- VPC: my-msk-cluster-vpc
- Subnets: my-msk-cluster-1a my-msk-cluster-1b my-msk-cluster-1c
MSK Configuration
- Wildcard DNS pattern: *.aklivity.[...].amazonaws.com *1
- Port number: 9094
- Private Certificate Authority: <private certificate authority ARN> *2a
Secure Public Access Configuration
- Instance count: 2
- Instance type: t3.small *3
- Role: aklivity-zilla-proxy
- Security Groups: my-zilla-proxy
- Secrets Manager Secret ARN: <TLS certificate private key secret ARN> *3
- Public Wildcard DNS: *.example.aklivity.io *4
- Public Port: 9094
- Private Certificate Authority: <private certificate authority ARN> *2
- Key pair for SSH access: my-key-pair *5
*Configuration Reference
1. Follow the Lookup MSK Server Names guide to discover the wildcard DNS pattern for your MSK cluster.
2. These can be the same Private Certificate Authority that authorizes existing clients connecting directly to MSK, allowing existing trusted client certificates to connect via Zilla proxy.
3. Consider the network throughput characteristics of the AWS instance type as that will impact the upper bound on network performance.
4. Replace with your own custom wildcard DNS pattern.
5. Follow the Create Key Pair guide to create a new key pair used when launching EC2 instances with SSH access.

Step 3. Configure stack options: `(use defaults)`

Step 4. Review

Confirm the stack details are correct and Submit to start the CloudFormation deploy.

Info

When your Zilla proxy is ready, the CloudFormation consoleopen in new window will show CREATE_COMPLETE for the newly created stack.

Verify Zilla proxy Service

This checks that the services and networking were properly configured.

Navigate to the EC2 running instances dashboard.open in new window

Check your selected region

Make sure you have selected the desired region, ex: US East (N. Virginia) us-east-1.

Select either of the Zilla proxies launched by the CloudFormation template to show the details.

Info

They each have an IAM Role name starting with aklivity-zilla-proxy.

Find the Public IPv4 Address and then SSH into the instance.

ssh -i ~/.ssh/<key-pair.cer> ec2-user@<instance-public-ip-address>

After logging in via SSH, check the status of the zilla-plus system service.

Service is running

Verify that the zilla-plus service is active and logging output similar to that shown below.

systemctl status zilla-plus.service

zilla-plus.service - Zilla Plus
   Loaded: loaded (/etc/systemd/system/zilla-plus.service; enabled; vendor preset: disabled)
   Active: active (running) since...

Check Ports

Check for the active ports with netstat.

netstat -ntlp

tcp6    0    0 :::9092    :::*    LISTEN    1726/.zpm/image/bin

Check Zilla Logs

You can get an stdout dump of the zilla-plus.service using journalctl.

journalctl -e -u zilla-plus.service | tee -a /tmp/zilla.log

systemd[1]: Started zilla-plus.service - Zilla Plus.
...

Check Cloud Init Logs

All output from cloud-init is captured by default to /var/log/cloud-init-output.log. There shouldn't be any errors in this log.

cat /var/log/cloud-init-output.log

Cloud-init v. 22.2.2 running 'init'...

Check the networking of the Zilla proxy instances to MSK.

DNS resolving

Verify that the instance can resolve the private Route53 DNS address.

nslookup *.aklivity.[...].amazonaws.com

Server:		***
Address:	***

Non-authoritative answer:
Name:	*.aklivity.[...].amazonaws.com
Address: ***

Check Ports

Check the communication over necessary ports with netcat.

nc -vz *.aklivity.[...].amazonaws.com 9094

Connection to *.aklivity.[...].amazonaws.com port 9094 [tcp/italk] succeeded!

Repeat these steps for each of the other Zilla proxies launched by the CloudFormation template if necessary.

Configure Global DNS

This ensures that any new Kafka brokers added to the cluster can still be reached via the Zilla proxy.

When using a wildcard DNS name for your own domain, such as *.example.aklivity.io then the DNS entries are setup in your DNS provider.

Navigate to the CloudFormation consoleopen in new window. Then select the my-zilla-proxy stack to show the details.

In the stack Outputs tab, find the public DNS name of the NetworkLoadBalancer. You need to create a CNAME record mapping your public DNS wildcard pattern to the public DNS name of the Network Load Balancer.

Info

You might prefer to use an Elastic IP address for each NLB public subnet, providing DNS targets for your CNAME record that can remain stable even after restarting the stack.

For testing purposes you can edit your local /etc/hosts file instead of updating your DNS provider.

Verify Kafka Client Connectivity

To verify that we have successfully enabled public internet connectivity to our Kafka cluster from the local development environment, we will use a generic Kafka client to create a topic, publish messages and then subscribe to receive these messages from our Kafka cluster via the public internet.

Install the Kafka Client

First, we must install a Java runtime that can be used by the Kafka client.

sudo yum install java-1.8.0

Now we are ready to install the Kafka client:

wget https://archive.apache.org/dist/kafka/2.8.0/kafka_2.13-2.8.0.tgz
tar -xzf kafka_2.13-2.8.0.tgz
cd kafka_2.13-2.8.0

Tips

We use a generic Kafka client here, however the setup for any Kafka client, including KaDeckopen in new window, Conduktoropen in new window, and akhq.ioopen in new window will be largely similar. With the Zilla proxy you can use these GUI Kafka clients to configure and monitor your Kafka applications, clusters and streams.

Configure the Kafka Client

With the Kaka client now installed we are ready to configure it and point it at the Zilla proxy.

We need to import the trusted client certificate and corresponding private key into the local key store used by the Kafka client when connecting to the Zilla proxy.

openssl pkcs12 -export -in client-1.cert.pem -inkey client-1.pkcs8.key.pem -out client-1.p12 -name client-1
keytool -importkeystore -destkeystore /tmp/kafka.client.keystore.jks -deststorepass generated -srckeystore client-1.p12 -srcstoretype PKCS12 -srcstorepass generated -alias client-1

In this example, we are importing a private key and certificate with Common Name client-1 signed by a private certificate authority. First the private key and signed certificate are converted into a p12 formatted key store.

Then the key store is converted to /tmp/kafka.client.keystore.jks in JKS format. When prompted, use a consistent password for each command. We use the password generated to illustrate these steps.

The Zilla proxy relies on TLS so we need to create a file called client.properties that tells the Kafka client to use SSL as the security protocol and to specify the key store containing authorized client certificates.

client.properties

security.protocol=SSL
ssl.keystore.location=/tmp/kafka.client.keystore.jks
ssl.keystore.password=generated

The password configured in client.properties should match the password used in the commands above used to create the key store.

Tips

As the TLS certificate is signed by a globally trusted certificate authority, there is no need to configure your Kafka client to override the trusted certificate authorities.

Test the Kafka Client

This verifies internet connectivity to your MSK cluster via Zilla Plus for Amazon MSK.

We can now verify that the Kafka client can successfully communicate with your MSK cluster via the internet from your local development environment to create a topic, then publish and subscribe to the same topic.

If using the wildcard DNS pattern *.example.aklivity.io, then we use the following as TLS bootstrap server names for the Kafka client:

b-1.example.aklivity.io:9094,b-2.example.aklivity.io:9094,b-3.example.aklivity.io:9094

Warning

Replace these TLS bootstrap server names accordingly for your own custom wildcard DNS pattern.

Create a Topic

Use the Kafka client to create a topic called zilla-proxy-test, replacing <tls-bootstrap-server-names> in the command below with the TLS proxy names of your Zilla proxy:

bin/kafka-topics.sh --create --topic zilla-proxy-test --partitions 3 --replication-factor 3 --command-config client.properties --bootstrap-server <tls-bootstrap-server-names>

A quick summary of what just happened

The Kafka client with access to the public internet issued a request to create a new topic
This request was directed to the internet-facing Network Load Balancer
The Network Load Balancer forwarded the request to the Zilla proxy
The Zilla proxy verified the client identity of the Kafka client
The Zilla proxy selected a matching client certificate to propagate client identity
The Zilla proxy routed the request to the appropriate MSK broker
The topic was created in the MSK broker
Public access was verified, authorized by trusted client certificate

Publish messages

Publish two messages to the newly created topic via the following producer command:

bin/kafka-console-producer.sh --topic zilla-proxy-test --producer.config client.properties --broker-list <tls-bootstrap-server-names>

A prompt will appear for you to type in the messages:

>This is my first event
>This is my second event

Receive messages

Read these messages back via the following consumer command:

bin/kafka-console-consumer.sh --topic zilla-proxy-test --from-beginning --consumer.config client.properties --bootstrap-server <tls-bootstrap-server-names>

You should see the This is my first event and This is my second event messages.

This is my first event
This is my second event

Conclusion

You have successfully deployed the Zilla Plus for Amazon MSKopen in new window Secure Public Access. Instructions on how to Monitor and Upgrade your Zilla proxy can be found in the managing a cloudformation stack section.