Keycloak deployment in EC2 | Adaltas

Why use Keycloak

Keycloak is an open-source identity provider (IdP) using single sign-on (SSO). An IdP is a tool to create, maintain, and manage identity information for principals and to provide authentication services for applications in a distributed network. Keycloak allows you to create multiple security policy domains called realms to manage objects such as users, applications, roles, and groups. It also enables you to register multiple users for the same application. Keycloak is either configured through a user interface called Admin Console or CLI commands.

A non-exhaustive list of use-cases includes:

  • Integration in a Kubernetes Cluster for authentication
  • Set up Keycloak for enabling login with social networks
  • Set up Keycloak as identity broker for another IdP

This article explains how to set up Keycloak in an AWS EC2 instance and make the identity provider publicly available through the internet.

1 Set up the EC2 instance

The main goal of the article is to know how to set up Keycloak in an AWS EC2 instance and not how to launch an EC2 instance on AWS. However, three different methods to configure and launch the EC2 instance are described in this article to adapt to the readers’ comfort and to specifically show which resources are required. If you are only interested in setting up the resources as fast as possible and want the scripts for idempotent reproduction, we recommend you skip parts 1.1 and 1.2 and go directly to part 1.3.

1.1 Use the Web Interface

1.1.1 Create the EC2 instance the first time

Launching the EC2 instance through the web interface is the user-friendliest solution.

You can also follow the official documentation on how to create an AWS EC2 instance. However, the following points provide the instance configuration steps.

  1. Log into your AWS account.

  2. Go to the region which suits you best, normally the closest to you.

  3. In the upper task bar type EC2 and go to this service.

  4. Click on Launch Instance.

  5. Choose Launch Instance again.

    EC2 Dashboard

    Keycloak in a Docker container does not require a lot of resources. Therefore, we choose a T2 micro instance with the minimum disk space. Installing Docker on an Amazon Linux AMI is very easy and we therefore use it.

  6. Give the instance a name.

  7. In Application and OS Images (Amazon Machine Image) choose the free tier Amazon Linux 2 Kernel 5.10 AMI 2.0.20221103.3 x86_64 HVM gp2 AMI.

    EC2 launch instance

  8. Choose the free tier t2.micro type in “Instance type”.

  9. If you already have a key pair choose it in the interface otherwise create one and click on the type RSA.

  10. In network settings SSH port 22 which is there by default allows us to log into the instance from another machine. However, Keycloak hosted on our EC2 instance needs access to the public internet, we therefore open either port 8080 and/or port 8443 with the TCP protocol. Set the source to Anywhere; this allows all users outside the virtual private network (VPC) to access the EC2 instance. Click on Edit, then Add security group rule and add the details mentioned previously.

EC2 network settings

EC2 add security group

EC2 configure port

  1. Leave the storage how it is to 8 GB.
  2. In Advanced details, there is the possibility to request a spot instance for cost saving, however, keep in mind that AWS can terminate the instance at any moment. In this same section under User data add the following code to install Docker and Postgres at launch. The reason why we install Postgres will be explained later on.
sudo yum update -y
sudo amazon-linux-extras install docker
sudo service docker start
sudo amazon-linux-extras enable postgresql13
sudo yum install postgresql postgresql-server -y
sudo postgresql-setup initdb
sudo systemctl start postgresql
sudo systemctl enable postgresql

EC2 add user data

If you forget to paste this code in User data before launching the instance, you will still be able to run each code line individually in the created EC2 instance afterwards.

1.1.2 Create and use a launch template

Create a launch template from the previously created instance to easily reproduce the set-up.

  1. As shown in the image, click on the instance.

  2. Then click on Actions.

  3. And finally click on Create Template from Instance.

    Create launch template

The launch template has been created and next time launch the instance from this template by clicking on Launch instance from template in the EC2 dashboard page as shown in the second image of the article.

1.2 Use the AWS CLI commands

If you have AWS CLI installed on your PC and are comfortable with the CLI commands, you may use them to set up the instance.

1.2.1 Set up the instance the first time

  1. Log on your AWS account with the following command and then give your credentials as explained in the official documentation

    Give your access key and private access key, then set your default region and choose json as the default output format.

  2. Create a key for the EC2 instance in your default region if you do not have one yet.

    aws ec2 create-key-pair \
      --key-name <your_key_name> \
      --query 'KeyMaterial' \
      --output text > your_key_name.pem
  3. Get your default VPC Id.

    aws ec2 describe-vpcs | grep VpcId
  4. Create a security group.

    aws ec2 create-security-group \
      --group-name <security-group-name> \
      --description "test the cli security group" \
      --vpc-id <your-vpc-id>
  5. Open ports 22, and do the same command for port 8080 and/or port 8443 by just replacing the port number.

    aws ec2 authorize-security-group-ingress \
      --group-id <sg-id> \
      --protocol tcp --port 22 --cidr
    aws ec2 authorize-security-group-ingress \
      --group-id <sg-id> \
      --protocol tcp --port 8080 --cidr
    aws ec2 authorize-security-group-ingress \
      --group-id <sg-id> \
      --protocol tcp --port 8443 --cidr
  1. Launch an EC2 instance of type T2 micro with an Amazon Linux AMI and with the Docker installation bash script. Paste the Docker installation code in a bash script and name it the same as in user-data of the run-instance command.

    aws ec2 run-instances --image-id ami-01cae1550c0adea9c \
      --count 1 --instance-type t2.micro \
      --key-name <your-keyname> \
      --security-group-ids <your-sg-id> \
      --user-data file://<docker-install-file-name>
  2. The following command creates a tag for the instance even though it is not mandatory.

    aws ec2 create-tags --resources <InstanceID> --tags Key=Name,Value=Keycloak

    To terminate the instance, use the following command:

    aws ec2 terminate-instances --instance-ids <instance-id>

1.2.2 Using a launch template for reproducibility

Let’s put in a template file all these settings from the existing instance that has been created to reproduce the same setup each time we launch the instance.

  1. First get the instance Id with the following command.

    aws ec2 describe-instances | grep "InstanceId"

    Note that if you have several instances running at the same time, you will have several Ids. You should therefore exclude the | grep "InstanceId" from the command and, in the long description, look which instance is concerned.

  2. Now extract all information to a json file by replacing the instance-id.

    aws ec2 get-launch-template-data \
      --instance-id <your-instance-id> \
      --query "LaunchTemplateData" >> instance-data.json
  3. With the obtained file, create a launch template.

    aws ec2 create-launch-template \
      --launch-template-name TemplateForWebServer \
      --version-description Version1 \
      --tag-specifications 'ResourceType=launch-   template,Tags=[Key=purpose,Value=Keycloak]' \
      --launch-template-data file://instance-data.json

Text in json format will appear on the console with the launch template Id.

If the launch template has already been created, get its Id with the following command.

aws ec2 describe-launch-template-versions \
  --versions "$Latest,$Default"
  1. Once we have the launch template and its Id, launch the EC2 instance as follows

    aws ec2 run-instances \
      --launch-template LaunchTemplateId=<launch-template-id>

1.3 Create the EC2 instance with Terraform

Another way to create all resources from scratch is to use Terraform. Terraform destroys all created resources with a simple command. In this case, it not only deletes the EC2 instance, but also the security group. The comfort of the Terraform scripts below is that they are applicable to all AWS users and therefore not customized for a specific one.

Keep in mind that the only element that the scripts below do not create is a new AWS SSH key for the instance. If you do not have one yet, please refer to the previous AWS CLI command or go in the web interface to create one and make sure it is created in the same region where you want to launch the EC2 instance.

  1. Check if Terraform is installed, if not, find the instructions on the hashicorp website and verify that the version is superior to 1.2.8 with the command terraform version on the command line.

  2. Create a file and paste the following code

      required_version = ">= 1.2.8"
    resource "aws_default_vpc" "default" 
      tags = 
        Name = "Default VPC"
    resource "aws_security_group" "sg" 
      name        = var.security_group_name
      description = "Allow inbound traffic on port 8080 and 8443"
      vpc_id      =
        description      = "open ssh port"
        from_port        = 22
        to_port          = 22
        protocol         = "tcp"
        cidr_blocks      = [""]
        ipv6_cidr_blocks = ["::/0"]
        description      = "open port 8080"
        from_port        = 8080
        to_port          = 8080
        protocol         = "tcp"
        cidr_blocks      = [""]
        ipv6_cidr_blocks = ["::/0"]
        description      = "open port 8443"
        from_port        = 8443
        to_port          = 8443
        protocol         = "tcp"
        cidr_blocks      = [""]
        ipv6_cidr_blocks = ["::/0"]
        from_port        = 0
        to_port          = 0
        protocol         = "-1"
        cidr_blocks      = [""]
        ipv6_cidr_blocks = ["::/0"]
      tags = 
        Name = var.security_group_name
    resource "aws_instance" "ec2" 
      ami           = "ami-01cae1550c0adea9c"
      instance_type = "t2.micro"
      vpc_security_group_ids = []
      user_data = <<-EOF
                  sudo yum update -y
                  sudo amazon-linux-extras install docker
                  sudo service docker start
                  sudo amazon-linux-extras enable postgresql13
                  sudo yum install postgresql postgresql-server -y
                  sudo postgresql-setup initdb
                  sudo systemctl start postgresql
                  sudo systemctl enable postgresql
      tags = 
        Name = var.aws_instance_name
    output "instance_ip" 
      value       = "$aws_instance.ec2.public_ip"
      description = "Public IP address"
    output "instance_dns" 
      value       = "$aws_instance.ec2.public_dns"
      description = "Public DNS"
  3. Now create a file and paste the following code. Type the AWS region where the resources should be created and where the SSH key is located

    provider "aws" 
      region = "<aws-region>"
  4. Finally create a file, paste the following code and replace the variable names

    variable "keyname" 
      type = string
      default = "<your-keyname>"
    variable "security_group_name" 
      type = string
      default = "<your-security-group-name>"
    variable "aws_instance_name" 
      type = string
      default = "<your-instance-name>"
  5. Once the 3 files have been created let’s initialize the environment

  6. Let’s check that no errors have been written in the files

  7. Now launch the set up

  • Use the terraform refresh command to get the information of the created resources again.
  • Apply the terraform destroy command to destroy all resources created by terraform for the environment.

2 Log into the EC2 instance

  1. We either get the public DNS name or the public IP address of the instance outputted by the terraform script. Alternatively, apply the following command to get the public DNS

    aws ec2 describe-instances | grep "PublicDnsName"

If you have several ec2 instances, leave out the | grep "PublicDnsName" and search the information corresponding to your EC2 instance in the long json script which is shown on the screen.

  1. If necessary ensure that the SSH key is not publicly viewable

  2. Log in the instance through SSH, it may take a few minutes for the instance to launch.

    ssh -i "<key-name>" ec2-user@<Public-DNS-Name or Public-IP>

3 Create a Docker container with Keycloak

3.1 Demo example in development mode

Now we create a Docker container running Keycloak publicly for a demo purpose.

  1. To avoid the sudo command for Docker, type the code below, exit your instance and log back in again

    sudo usermod -aG docker $USER
  2. Now let’s pull the image. At the time of writing the latest version is 20.0.3.

    docker pull
  3. Run the container by enabling port 8080 inside the container and on the instance.

    export KEYCLOAK_DEV_ADMIN_PASSWORD=<your password>
    docker run -d \
      --name keycloak-dev \
      -p 8080:8080 \
      -e KEYCLOAK_ADMIN=admin \
  4. If we type docker ps we see that the container is running on port 8080.

Docker running container

  1. Open the web browser with [http://ec2-public-ip:8080] to access the page below.

Keycloak entry page

However, if we click on Administration Console we cannot log in because the data flow between your EC2 instance and the public internet is by default protected with SSL encryption in Keycloak.

Keycloak HTTPS required

And adding https:// in front of the public IP only gives an error as follows

Keycloak secure connection failed

If we connect with a machine inside our VPC, we will have access to the console. In this example, we disable the SSL/TLS requirement to have access to Keycloak through the public internet.

  1. To deactivate SSL/TLS let’s enter the container in interactive mode.

    docker exec -it keycloak-dev bash 
  2. In the container set the path to use the script.

    export PATH=$PATH:/opt/keycloak/bin
  3. Then run the following command to set the server and the user of the master realm. config credentials --server http://localhost:8080/ \
      --realm master \
      --user admin --password <your password>
  4. By typing the following command, we get the realm details outputted in json format and we see that the value for sslRequired is set to external at the bottom of the picture. This indicates that SSL/TLS is required outside of our VPC. get realms/master

Realm details

  1. Disable the SSL/TLS requirement for the master realm. update realms/master -s enabled=true -s sslRequired=none
  2. Now on the browser click on “Administration Console” again and this time the login page appears.

Keycloak login page

  1. Give the credentials and obtain the following page.

Keycloak master realm UI

  1. If we go in the master realm to Realm settings,the tab Require SSL is now set to None.

Keycloak realm settings

Keep in mind that disabling the SSL/TLS requirement is not recommended in production mode and/or if you deal with sensitive data.

Before turning to the next section, stop the container with the command docker stop keycloak-dev since the new container will require the same host port. Delete it with docker rm keycloak-dev as it is not being used any longer.

3.2 Deploy Keycloak in production mode with a self-signed SSL/TLS certificate

You might wonder why we have installed Postgres with the user data since we did not use it yet. Keycloak comes with an in-memory database called H2. However, it is insufficient for large data and therefore not recommended for use in production mode. That is the reason why we installed an external database to connect to Keycloak. In this case we use Postgres, however you have the choice to connect other databases such as Oracle or MySQL.

If you have not done the demo example in development mode, perform step 1 and 2 of that section and continue from here on.

  1. Check that Postgres is working

    sudo systemctl status postgresql

Postgres status check

  1. Postgres has already a user created called postgres. Let’s set the database password and then type exit to leave.

    sudo su - postgres
    psql -c "ALTER USER postgres WITH PASSWORD '<your password>';"
  2. Open file /var/lib/pgsql/data/pg_hba.conf with nano or vim as root user. Under section # IPv4 local connections:, comment out host all all ident and add host all postgres md5. This action enables only the user postgres on the localhost to access the database with a password. Postgres has no other purpose than being used by Keycloak in our case. You can find more about authentication methods on the Postgres website.

    Postgress configuration file

  3. Restart Postgres.

    sudo systemctl restart postgresql
  4. Create a folder for the SSL certificats.

    mkdir certificates
    cd certificates
  5. Create the self-signed certificat and the key in pem format.

    openssl req -newkey rsa:2048 -new -nodes -x509 -days 3650 -keyout key.pem -out    cert.pem
  6. Set passwords and hostname. The database password must be same as the one you have set previously in Postgres.

    export KEYCLOAK_ADMIN_PASSWORD=<your keycloak admin password>
    export DB_PASSWORD=<your db password>
    export KEYCLOAK_HOSTNAME=<your public ip>:8443
  7. Create the container with the following command.

    docker run -d  \
      --name keycloak-prod \
      -v /home/ec2-user/certificates:/certificates \
      --network=host \
      -e KEYCLOAK_ADMIN=admin \
      start \
      --features=token-exchange \
      --https-certificate-file=/certificates/cert.pem \
      --https-certificate-key-file=/certificates/key.pem \
      --hostname=$KEYCLOAK_HOSTNAME \
      --proxy=edge \
      --db=postgres \
      --db-url=jdbc:postgres://localhost:5432/postgres \
      --db-username=postgres \

    Here we mount a volume with the certificates to the container. Keycloak then stores them in a keystore with the options --https-certificate-file and --https-certificate-key. More about configuring TLS is on the Keycoak website.

    We order the container to use the same network as the host with --network=host. Therefore we do not have to do any network configuration on the container. Keycloak uses by default port 8080 and 8443 for respectively http and https. Postgres is connected through its default port 5432.

    Here, we set the public IP address as hostname as we want the Admin Console to be viewable on the public internet. With the option --proxy=edge, our server is running behind a TLS termination proxy. Read the official documentation for further information on this topic.

    The last options in the command are mandatory to connect the Postgres database installed on the EC2 instance to the container.

  8. After about 30 seconds, open the web browser with [https://ec2-public-ip:8443] to access the page below in Firefox

    Browser security warning for HTTPS

  9. Since we used a self-signed certificate we get a warning message. In Firefox, click on Advanced and accept the risk and the browser links you then to the same console seen above.

Accept self-signed certificate

Now that we have access to the Admin Console, we can use it for any further operations such as creating clients or users, which is explained in the Keycloak documentation.

Using the CLI commands

The use of the CLI commands first necessitates the configuration of the truststore.

  1. Therefore, enter the container like in the previous section.

    docker exec -it keycloak-prod bash 
  2. Set the path for the executable files.

    export PATH=$PATH:/opt/keycloak/bin
  3. Create a truststore.jks file with your pem certificate and then chose a password.

    cd /opt/keycloak
    keytool -import -alias root -keystore truststore.jks -file /certificates/cert.pem
  4. Configure the truststore to use the truststore.jks file config truststore --trustpass <trustsore password> ~/truststore.jks

From now on, using further CLI commands does not differ from the demo example in development mode. You can find more CLI commands in the official documentation.

We have finally setup Keycloak in a Docker container on an EC2 instance and it is ready for any further operations.

Do not forget to terminate all your created resources once they are no longer being needed.

Luis Robinson

Next Post

How To Hang Golf Towel With Hole in Middle

Sun Mar 26 , 2023
“The Art of Hanging a Golf Towel with a Hole in the Middle: A Guide for the Amused Golfer”   Golfing is not just a sport, it’s an art form. From the perfect swing to the precise aim, every aspect of the game requires finesse and skill. But let’s not […]
How To Hang Golf Towel With Hole in Middle

You May Like