CDK construct library that allows you to create KeyCloak on AWS in TypeScript or Python
Note
This project has been migrated to CDK v2.
CDK v1 compatible version is deprecated now.
import { KeyCloak } from 'cdk-keycloak';
const app = new cdk.App();
const env = {
region: process.env.CDK_DEFAULT_REGION,
account: process.env.CDK_DEFAULT_ACCOUNT,
};
const stack = new cdk.Stack(app, 'keycloak-demo', { env });
new KeyCloak(stack, 'KeyCloak', {
certificateArn: 'arn:aws:acm:us-east-1:123456789012:certificate/293cf875-ca98-4c2e-a797-e1cf6df2553c',
keycloakVersion,
});
Use keycloakVersion
to specify the version.
new KeyCloak(stack, 'KeyCloak', {
certificateArn,
keycloakVersion: KeycloakVersion.V15_0_2,
});
To specify any other verion not defined in the construct, use KeycloakVersion.of('x.x.x')
. This allows you to specify any new version as soon as it's available. However, as new versions will not always be tested and validated with this construct library, make sure you fully backup and test before you use any new version in the production environment.
The KeyCloak
construct provisions the Amaozn RDS cluster for MySQL with 2 database instances under the hood, to opt in Amazon Aurora Serverless, use auroraServerless
to opt in Amazon Aurora Serverless cluster. Please note only some regions are supported, check Supported features in Amazon Aurora by AWS Region and Aurora DB engine for availability.
// Aurora Serverless v1
new KeyCloak(stack, 'KeyCloak', {
certificateArn,
auroraServerless: true,
keycloakVersion,
});
// Aurora Serverless v2
new KeyCloak(stack, 'KeyCloak', {
certificateArn,
auroraServerlessV2: true,
keycloakVersion,
});
Behind the scene, a default RDS cluster for MySQL with 2 database instances will be created.
To create single RDS instance for your testing or development environment, use singleDbInstance
to turn on the
single db instance deployment.
Plesae note this is not recommended for production environment.
new KeyCloak(stack, 'KeyCloak', {
certificateArn,
singleDbInstance: true,
keycloakVersion,
});
Define autoScaleTask
for the ecs service task autoscaling. For example:
new KeyCloak(stack, 'KeyCloak', {
auroraServerless: true,
nodeCount: 2,
autoScaleTask: {
min: 2,
max: 10,
targetCpuUtilization: 60,
},
});
You can deploy the workload in the existing Vpc and subnets. The publicSubnets
are for the ALB, privateSubnets
for the keycloak container tasks and databaseSubnets
for the database.
The best practice is to specify isolated subnets for databaseSubnets
, however, in some cases might have no existing isolates subnets then the private subnets are also acceptable.
Consider the sample below:
new KeyCloak(stack, 'KeyCloak', {
certificateArn: 'arn:aws:acm:us-east-1:123456789012:certificate/293cf875-ca98-4c2e-a797-e1cf6df2553c',
vpc: ec2.Vpc.fromLookup(stack, 'Vpc', { vpcId: 'vpc-0417e46d' }),
publicSubnets: {
subnets: [
ec2.Subnet.fromSubnetId(stack, 'pub-1a', 'subnet-5bbe7b32'),
ec2.Subnet.fromSubnetId(stack, 'pub-1b', 'subnet-0428367c'),
ec2.Subnet.fromSubnetId(stack, 'pub-1c', 'subnet-1586a75f'),
],
},
privateSubnets: {
subnets: [
ec2.Subnet.fromSubnetId(stack, 'priv-1a', 'subnet-0e9460dbcfc4cf6ee'),
ec2.Subnet.fromSubnetId(stack, 'priv-1b', 'subnet-0562f666bdf5c29af'),
ec2.Subnet.fromSubnetId(stack, 'priv-1c', 'subnet-00ab15c0022872f06'),
],
},
databaseSubnets: {
subnets: [
ec2.Subnet.fromSubnetId(stack, 'db-1a', 'subnet-0e9460dbcfc4cf6ee'),
ec2.Subnet.fromSubnetId(stack, 'db-1b', 'subnet-0562f666bdf5c29af'),
ec2.Subnet.fromSubnetId(stack, 'db-1c', 'subnet-00ab15c0022872f06'),
],
},
});
This library support AWS China regions cn-north-1
and cn-northwest-1
and will auto select local docker image mirror to accelerate the image pulling. You don't have to do anything.
See CONTRIBUTING for more information.
This project is licensed under the Apache-2.0 License.