🚀Deployment

A guide on how to deploy services to both production and staging.

Deploy to staging

To deploy to staging, merge to the staging-main branch and push the code. The deployment will go through automatically! You can see the GitHub staging deployment actions to check the status. You will also get a notification in the deploys-staging Slack channel once it has gone through.

• Run migrations if needed.

• Update the environment variables if needed.

Deploy to production

The production deployment is done with a GitHub action. You can select which services you want to deploy and the branch (it should 99% of the time be main), and click Deploy!

• Run migrations if needed.

• Update the environment variables if needed.

Updating environment variables

Environment variables are kept in the kube repo. There is a folder for production and one for staging. There we have environment variables for frontend and backend.

To push an update to the Kubernetes cluster:

1. Make sure you have an up-to-date version of the Kube repo

2. Make the changes you want, and commit and push them! Otherwise, they could be overwritten by a colleague.

3. Run the [production/staging]./update-config-map script to update the values in the cluster.

4. Re-deploy the necessary services.

Running migrations

Running database migrations is, for now, not a part of the automated deployment. They need to be run manually. To run migrations:

1. Make sure you are in the appropriate branch. If you are running them on production, this is main. For staging, staging-main. Make sure the branch is up-to-date.

2. Export the directory where crowd.dev is in your system. For example export CROWD_CHECKOUT_DIR=/Users/joanreyero/Documents/crowd.dev.

3. Run the script ./migrate-up.sh to run the migrations.