AWS Marketplace install
The Hazy platform can be purchased and deployed through the AWS Marketplace as an AMI (Amazon Machine Image) server offering. This allows customers to deploy the software on their own AWS private cloud infrastructure.
Deployment¶
After purchasing the offer, you will be provided with access to the Hazy AMI. To begin using the software, you will need to launch the machine image as an EC2 instance. This can be achieved through the Launch wizard or AWS API. There are a few settings to take note of when launching this instance:
Instance type¶
The instance should be appropriately sized to run training and generation on the datasets you wish to synthesize. A good starting instance to trial the software would be an r5.xlarge (4CPU, 32GB ram). For more detail refer to Hardware Requirements.
Keypair¶
You will need to select a SSH keypair to allow access to the machine for the purposes of maintenance and administration. The SSH username for this AMI image is ubuntu.
Networking¶
By default, the Hazy platform runs on port 8000. This HTTP endpoint allows access to the Hazy Hub. If you need secure access over the public internet, there are a few options:
- SSH port forwarding (locally and remotely to port 8000)
- HTTPS via an Amazon Elastic Load Balancer or other TLS termination proxy
It is highly recommended that SSH connections are enabled for the machine for the purposes of administration.
Storage¶
As part of provisioning, you should choose an appropriately sized instance disk to run training and generation. As a minimum, we recommend provisioning an instance disk of no less than 30GB. With additional configuration, it is possible to use EBS (Elastic Block Store) volumes as the backing store for the Hazy software and data.
Instance role¶
You will need to create an IAM Role and select it when running the AMI solution. This role must permit the instance to make License Manager CheckoutLicense calls to AWS. This is required to enable reading of your license information for the Hazy software.
Additionally, should you require access to source datasets on AWS S3, this role should permit access to the required buckets in your AWS account.
Launch the instance¶
Once configuration is completed, you are ready to launch your Hazy instance. Please allow approximately 1 minute for the instance and software to start up.
Getting started¶
Logging in¶
Once your Hazy instance has started, it’s time to log in for the first time. To connect, navigate in your web browser to:
http://<instance_ip_adress>:8000/
You will be prompted for a username and password. The username should be set to hazy and the initial password will be the Instance ID of the instance used to launch the software. Once logged in you will be prompted to change this password.
You will now be logged into your deployed Hazy instance and ready to begin training. Please refer to the User Guide to begin generating synthetic data.
Maintenance and administration¶
To perform maintenance/configuration changes you will need to SSH into the Hazy instance.
To stop the Hazy Hub run:
systemctl stop hazy
Configuration and environment variables for the Hazy software can be found in /app/hazy.yaml. For more detailed information on these settings refer to the Installation Guide.
Once any changes have been made, you can start the Hazy Hub with:
systemctl start hazy