- Run the following steps to setup the project initially.
- Update the git config to point git hooks to ./git_hooks by running
git config core.hooksPath git_hooks. This allows us to automatically run Docker and Kubernetes deployment hooks upon a git push. - Create a service account for the GCP project you are currently using and ensure it has access to read/write from all cloud buckets since this is required when the initial GCR bucket is created upon the first Docker image push to GCR.
- Download the service account key and place it under
service_acc_keys/service-acc-key.json - Authenticate gcloud using the service account key
gcloud auth activate-service-account [ACCOUNT] --key-file=service_acc_keys/service-acc-key.json - Install npm on your machine
- Install Docker on your machine
- Install Kubernetes (kubectl) on your machine
- Setup a Google Cloud Kubernetes cluster.
- Connect Kubernetes to a GCP Kubernetes cluster by running
gcloud container clusters get-credentials CLUSTER_NAME --region=COMPUTE_REGION - Setup a Google Cloud Redis MemoryStore instance.
- You can run automation scripts present in
automation/(which are also automatically run upon agit push origin main) to deploy changes onto the Kubernetes cluster - If you make any changes to the MongoDB StatefulSet, then you will have to follow the additional instructions present in
kubernetes/README.mdto setup the MongoDB ReplicaSet. - Currently react-frontend is unable to access the flask-server-service directly from the cluster when you test it on a local browser with port forwarding. Hence, for development, port-forward both the react-frontend-service as well as the flask-server-service and make API calls to
localhost:<PORT_OF_FLASK_SERVER_SERVICE>from within the react app. Use the scriptkubernetes/port-forward-services.shto port forward both the frontend and the server.
- Add some colored text to make it easier to know whats going on when the automation scripts are run
- Support authentication/authorization
- Add in a server side as well as a client side cache.