Dependencies

• docker & docker-compose

Build and run docker images

execute ./build_and_run_docker.sh

Tools

to execute get / post queries directly on back/ml: https://www.postman.com/

Setup a new cloud environment

• Create a new project on google cloud

Setup back

• Create a new service on Google cloud run
• Create a new spring boot config (application-{ENV}.yml)
• following prod environment, set appropriate environment variables for back container in cloud run (https://cloud.google.com/run/docs/configuring/environment-variables) (SPRING_PROFILES_ACTIVE, and variables used in application-{ENV}.yml like ML_SERVICE_URL...)
• Adapt deploy_to_cloud_run.sh (to correct project and cloud_run_service)

Setup front

• Create a new project on firebase. On creation, associate it to the google cloud project
• Activate firebase hosting
• Create a new npm build config (build:ENV in package.json) to set correct back api url
• Adapt to deploy_to_firebase_hosting.sh (to deploy to correct project firebase project)

Setup ml

• Create a new VM on Google compute engine
• Use compute engine OS customized for container (https://cloud.google.com/compute/docs/containers/deploying-containers)
• Accept connection by http (will be called by back)
• Set static public IP (to be seen by back container) (https://cloud.google.com/compute/docs/ip-addresses/reserve-static-external-ip-address), for now there is no way to access it internally (inside google cloud) from cloud run (https://cloud.google.com/run/docs/using-gcp-services)
• Connect to database (https://cloud.google.com/sql/docs/postgres/connect-compute-engine)
• Define the environment variables used by ml service in google cloud web console. The list of the variables is in the file docker-compose.yml
  • Open the VM instances screen & click on the appropriate VM to open the details: https://console.cloud.google.com/compute/instances?project=ibo-speak&instancessize=50
  • Click on Edit & then on 'Advanced container options' link
  • Add the variables within 'Environment variables' section
• adapt deploy_to_comput_engine.sh

Setup PostgreSQL database

• Create a Postgre 12 image on google cloud SQL
• Connect back cloud run to postgresql cloud SQL: https://cloud.google.com/sql/docs/postgres/connect-run
• Connect to database with PgAdmin (https://cloud.google.com/sql/docs/postgres/connect-admin-proxy) and run init scripts to create client user for back (change password!) and tables
• Connect cloud sql to ml service VM on compute engine (https://cloud.google.com/sql/docs/postgres/connect-compute-engine) with public IP