Deploying DocAI Workbench

Currently, the marketplace-guided installation for DocAI Workbench is not available. To deploy DocAI Workbench in your GCP project, you'll need to trigger a manual install. Follow the steps outlined below. You only need the marketplace license file to run the DocAI Workbench application.

Major Steps

• Create a new GCP project and associate it with your GCP billing account

• Download GCP Marketplace license

• Get the deployment

• Configure the Workbench deployment

• Install the Workbench

Create a new GCP project

• Go to the Google cloud console (cloud.google.com) and create a new GCP project.

• Make sure that the new project is associated with your billing account and it is not under evaluation. You will need to set it as a paid account

• When you are inside the project then make sure you have at least one VPC network. If you see that a default network is created by the Google cloud then you are all set. Otherwise create a network and name it as a default.

Download marketplace license

• Follow the steps provided in this online documentation link - https://trillo.gitbook.io/deploying-trillo-workbench-in-google-cloud

• Rename the downloaded file as license.yaml (For enterprise license, please contact [e-mail])

Get the deployment

• Contact [e-mail] to request installation file - deployment.zip

• Upload and unzip deployment.zip to your account under Google cloud shell.

Configure the DocAI Workbench

The deployment configuration is provided in the shell script env_global.sh

• You must provide the name of the server including the domain (eg: wb.domain)

• Once the installation is complete then update the DNS A records for the two names (eg: wb.domain and api.wb.domain)

• Change the region and zones if needed. Everything will be created in the specific zone and the region afterwards.

• If a default Google cloud project is created, there is no need to change the network and subnetwork configurations.

• The rest of it can be left as is for the successful installation of the workbench.

Variable Name	Value	Comment
APPSERVER_UI_NAME	docai.$APPSERVER_NAME	Set app server UI name to server name
APPSERVER_NAME	api.$APPSERVER_NAME	Set the app server name with the domain (eg: server.domain)
REGION		Set region
ZONE		Set primary zone
ZONE2		Set secondary zone
NETWORK_SHORT	default	Use default network name
SUBNETWORK_SHORT	default	Use default subnetwork name
OTHER_NETWORK_PROJECT	$PROJECT_ID	Use current project for network

Install the Workbench

• Make sure you are ssh-ed into the deployer machine

• Copy the license file from the home folder of the deployer machine to the deployment folder as follows

cd ~/deployment

cp ~/license.yaml ./4-gke

• cd to the deployment folder (as above) and run the script ./runall.sh

• The installation will start the scripts and run for more than 15 mins. It will provision the followings

  • GCP Services

  • Reserve IP addresses for the Workbench frontend and backend servers.

  • Create the bucket for file storage.

  • Provision MYSQL 8 server for data storage

  • Provision Redis for caching

  • Optionally provision nodes for secure file transfer.

  • Finally, install all microservices in the kubernetes cluster. It will provision the external load balancer for the frontend and backend servers with managed secure certificates.

• Once the installation is complete then the scripts and the configuration results are all copied to the newly provisioned cloud storage bucket. The configuration results include the default admin credentials.

• The load balancer will take around another 15+ minutes to stabilize but notice that it will not be able to retrieve valid security certificates until unless the DNS records are updated. Therefore proceed to update your DNS A records by looking at the IP address reserved for the Workbench frontend and backend servers. IP addresses are derived from the following page.

• Keep checking the URL (eg. wb.domain) inside the browser till the main login screen is observed. You can jump through the main URL and the login screen from the Google cloud console as below.

• This is how the login screen looks like. The default admin credentials are provided on the last screen once the installation is complete.

• Once you're inside the workbench then create a new user. Notice that the admin account is not meant to operate the workbench. The newly created users will be able to operate all functions of the workbench.

Additional notes

• For terraform templates, please contact [e-mail]

• The installation of DocAI Workbench include:

  • DocAI Manager

  • File Manager