Spring boot and Docker in Google Compute Engine

Before diving in 🙌

• Request access to the Google Cloud Developer Console and MongoDB Atlas console
• Install the latest stable version of Maven. This will be your main version across your machine.

Getting Started 🚀

This project contains two profiles:

• dev
• prod

To run the desired profile use mvn spring-boot:run -P dev|prod

Setup MongoDB

Setup Firebase Emulator locally

Follow the official Firebase guide to set up the emulator locally.

Creating Routes

Routes can be found in src/main/***/http/controller

Creating Mongo Documents and Repositories

Documents can be found in src/main/***/model Repositories can be found in src/main/***/repository

Logging

This app uses Logback for logging. You can find the config in /src/main/resources folder When running in prod Logs are sent to Cloud Logging via Logback Cloud logging plugin. The Logback plugin only reports logs from the LF4J logging API, so we only use the org.slf4j.Logger interface for logging. When running in dev profile, logs are configured to write to the console.

Use this setup to configure Docker to work with Cloud Logging when moving to a new VM

Secrets Management 🔒

Use Google Cloud Secrets Manager to manage secrets (API keys, passwords, database URLs, etc.)

Deployments

Deployments are currently automated with GitHub actions. The workflow file is located at /.github/workflows/deploy.yml The app runs on a Google Compute Engine VM with full GCP API permissions and required scopes.

Building the production docker image on a new machine

1. Install Docker and Gcloud CLI
2. Run gcloud auth application-default login to authenticate with Google Cloud Platform
3. Run CP $HOME/.config/gcloud/application_default_credentials.json cred.json from the project root folder.
4. Temporarily allow the production MongoDB Atlas to accept traffic from your local machine's IP address.
5. Run docker-compose up <service-name> from the project root folder. Use jvm or native as the service name.

Publishing a new version to Google Artifact Registry

You will need to have write access to our Google Artifact Registry on Google Cloud Platform and install docker on your machine.

1. Run gcloud auth configure-docker us-central1-docker.pkg.dev to enable Google Cloud CLI to authenticate requests to Artifact Registry.
2. Run the docker tag web <GCP_IMAGE_NAME> && docker push <GCP_IMAGE_NAME> command to publish the new version to Google Artifact Registry.

Updating the container image with the new image version

$ gcloud compute instances update-container [instance-name] --zone=[zone-name] --container-image=[container-image-name]

Replace the instance-name, zone-name and container-image-name with the appropriate values.

Inspecting the app on the VM

While you will not need to log into the VM to get Telementry information, you can SSH into the VM and run the following command to inspect the app

1. Run docker ps to view the docker instances running
2. Run docker attach [container-id] to attach to the container and start seeing logs. Note that you will only see logs from when you attach, not the past logs.

Use Cloud logging to inspect the logs and health of the machine.

Tests 🧪

Unit test

Integration test

The goal is to achieve close to production behaviour as possible. Database and Repositories are powered by an Embedded DB that only runs in test profile.

1. Mockito is used for Mocking external services
2. MockMvc is used for the Integration test

While mvn clean test is good for running the tests during development, we advise you use the following command to build the test.Dockerfile to verify that the test can run in an isolated environment without any preconfiguration on your local machine.

docker build -t test -f test.Dockerfile .

A successful build from the above command indicates that all tests are passing on your local machine.

---

Here are some tips on how to optimize java application.