Build Quality Microservice with gRPC and GitHub Workflows: Hands on Guide (Part 2)

The path to an amazing software supply chain. It only takes few minutes

Everything mentionned in this article is available on this repository https://github.com/chaisarfati/UserMicroservice

In the first part of this gRPC guide, we set up a repository containing a Python application that implements a gRPC service definition. We created a GitHub workflow (pipeline) that generates interface code from the protobuf, compiles the code, tests it, and commits the generated code to the repository.

In this second part, we’ll define a Dockerfile and add it to our repository. Then, we’ll add a job to our GitHub workflow that builds an image from this Dockerfile and pushes it to GitHub’s Container Registry.

Step 1: Defining the Dockerfile

In this Dockerfile:

• We start with a lightweight Python 3.9 image.
• We copy our service files into the container.
• We install all dependencies (including gRPC-related dependencies).
• We expose the gRPC port (50051 in this case) so that other services can communicate with it.
• Finally, we run our Python gRPC service using the command python server.py.

Step 2: Adding a Docker Build Job to the GitHub Workflow

Next, we’ll add a new job to our GitHub workflow that will build and push our Docker image to GitHub’s Container Registry. Here’s the updated workflow:

Breaking down the job:

1. Checkout the Code: The first step checks out the code from the current branch, which is determined by the branch name we defined in the init job.
2. Set Up Docker Buildx: This action sets up Docker Buildx, a tool for building Docker images that supports multi-platform builds.
3. Log in to GitHub Container Registry: This step logs in to GitHub’s Container Registry using the secret stored in GHPR_TOKEN, allowing us to push the image to GitHub.
4. Sanitize Repository and Branch Names: Here, we make sure the repository and branch names are lowercase and properly formatted to create valid Docker tags.
5. Build and Push the Docker Image: Finally, we build the Docker image using the Dockerfile we defined earlier, tag it using the branch name and short commit SHA, and push it to GitHub’s Container Registry.

Step 3: Verifying the Build

Once this pipeline is complete, you’ll see your Docker image in the GitHub Container Registry, tagged with the branch name and commit SHA such as ghcr.io/<repository_owner>/<repository_name>:<branch_name>.<commit_sha>.docker

If we go to the repository ‘Packages’ tag and we clear the filters, we should see this:

There you go. Ready to deploy your microservice on your favorite container managing platform

That’s all for this part. In the next article, we’ll move on to packaging the microservice for deployment using Helm, an increasingly popular tool for Kubernetes package management.

Stay tuned!