prod-deployment

prod-deployment

This is the deployment at https://www.xaidemo.de.

Production deployment of the XAI demonstrator on Google Cloud Platform. The frontend is served through Firebase Hosting with the backends deployed on Google Cloud Run.

Set Up

To set up the XAI demonstrator on the Google Cloud Platform follow these steps:

1. Create a GCP project and enable billing.
2. Activate Google's Cloud SDK.
3. Create a service account for the GitHub workflows and assign the following roles to it: Service Account User, Cloud Run Admin and Cloud Storage Admin.
4. Create a key for your workflow service account.
5. Create the following GitHub secrets:
   • GCP_PROD_EMAIL - the email adress of your workflow service account
   • GCP_PROD_PROJECT_ID - your project ID
   • GCP_PROD_REGION - the region you want your project to be hosted in
   • GCP_PROD_SA_KEY - the key your created in step 4
6. Activate Firebase Hosting for your GCP project (without Google Analytics)
7. To generate the Firebase token, install the Firebase CLI and log in using your Google account associated with your GCP project with firebase login:ci. Store the token as GCP_PROD_FIREBASE_TOKEN.
8. Set the project name in .firebaserc to the name of your GCP/Firebase project.

Be careful to insert the exact identifiers when creating the GitHub secrets. Common problems are extra spaces, inserting the computing zone (e.g. us-west1-a) instead of the region (us-west1) and copying only parts of the key json.

Now the XAI demonstrator is ready to be deployed to your GCP project by the Prod Deployment workflow!