Global Water Watch Service

Backend API for synchronous operations within Global Water Watch

Testing

docker-compose up
pytest

Local deployment

Currently, I have 2 ways of deploying locally

Debugging

For debugging, it is best to debug the main.py file in python, and use docker-compose to get the database running: docker-compose up

Get earthengine credentials: make get_secrets

And add the following environment variables (to your launch.json in VSCode) and .env:

{
    "EE_SA": "dagster-workloads@global-water-watch.iam.gserviceaccount.com",
    "EE_PK": "${workspaceFolder}/gcloud_dist/privatekey.json"
}

We can then use psql to create the postgis extension: psql -U testuser -h localhost -d testdb using the testpassword.

Then use (for example) VSCode to debug main.py.

Local kubernetes

To test kubernetes (later for some extra functionality like workflows), you can deploy a local cluster. This is simplified in the Makefile.

• Install kind.
• Install kubectl
• Install helm

make build-docker
make deploy-local

Check out the database: in one terminal:

export PGMASTER=$(kubectl get pods -o jsonpath={.items..metadata.name} -l application=spilo,cluster-name=gwwuser-gww,spilo-role=master -n gww)
kubectl port-forward $PGMASTER 6432:5432 -n gww

in another:

export PGPASSWORD=$(kubectl get secret -n gww gwwuser.gwwuser-gww.credentials.postgresql.acid.zalan.do -o 'jsonpath={.data.password}' | base64 -d)
export PGSSLMODE=require
psql -U gwwuser -h localhost -p 6432 -d gwwdb

Check out the api:

kubectl port-forward -n gww svc/gww-api 8080:80

After you are done:

kind delete clusters waterwatch

Note: This is to test working with other software, for debugging, using VSCode together with docker-compose and FastAPI's --reload option helps developing quickly.

Deploy on kubernetes

This repository contains a helm chart that deploys the API to kubernetes. For example values, check out the values file.