Have you ever been in a situation where your code worked perfectly fine in your local environment but broke in production? Great, this article will help you prevent being in that dire situation. Furthermore, this tutorial will help you create a streamlined microservice development workflow for Kubernetes.
Developing microservices in a local development environment can present several challenges for developers.
To overcome some of these challenges, organizations use tools and frameworks such as Tilt and Docker Desktop(Kubernetes) that streamline the local development process. These tools help automate many manual steps in setting up a local development environment, making it easier to focus on writing code and testing applications and not worry about the underlining infrastructure, cross-dependency, and software packaging.
Test it like it’s in production
Tilt: Development Environment as Code is a powerful tool that makes it easy to develop, build, and test applications in a Kubernetes environment. It automates the manual steps in setting up a development environment, such as building containers, uploading them to a registry, and deploying them to a cluster.
Some of the standout features include,
Docker Desktop provides an easy-to-use development environment for building, shipping, and running dockerized applications. Cross-platform compatibility, intuitive GUI, and docker-extensions make it a must-have on every developer toolkit.
Fun fact: Docker has acquired Tilt.
In this tutorial, we will leverage the capabilities of Tilt and Docker Desktop to build a streamlined microservice development workflow.
We have set up a sample repository that demonstrates the Dev Environment as Code workflow. Please give it a spin for hands-on exploration. You can also jump straight to tilt-internals to understand it better.
We have configured an example repository with Tilt to make it easy for you to try. You can get your microservices dev environment up and running with three commands.
git clone https://github.com/hivenetes/bots
tilt-resources/ to the root dir:
# Configurations of k8s contexts, service names, ports, etc. `tilt_config.json` is referenced in the `Tiltfile.`
cp tilt-resources/local/tilt_config_local.json tilt_config.json
# Open the UI on a browser at http://localhost:10350/
When you run
tilt up, Tilt looks for a file named
Tiltfile in the current directory, which defines your dev-environment-as-code. Next, you will be redirected to the Tilt-UI as shown below:
bots application is now accessible via
# Cleans up all the resources specified in the Tiltfile
Wait, what just happened?
Don’t worry. Let’s now dive into the internals to see what was going on!
A resource is a collection of workloads managed by Tilt. Some of the workloads include,
Here is an example of a resource definition.
A Titfile is a configuration file written in Starlark, a dialect of
python. It consists of all the resource definitions and how they connect. It’s actual code; you can use conditionals, loops, and functions to define your microservice environment.
# Example Tiltfile
# Read the configuration
if not os.path.exists("./tilt_config.json"):
fail("Config not found")
cfg = config.parse()
# Build each microservice image as stated in the tilt_config.json file
for microservice in cfg.get("microservices"):
When the Tiltfile is executed, the YAML is packaged as a resource and sent to the
Tilt engine is a controller that constantly watches the files fed into it. If Tilt detects any changes that might affect the output of the Tiltfile, it re-evaluates the Tiltfile.
The Tilt engine watches critical events like:
For example, as you edit your code, Tilt will automatically update steps such as building an updated container image, and if any resources contain Kubernetes objects, these end up automatically deploying to your development cluster.
Inner development loop is the iterative process of writing, building, and debugging code that a single developer performs before sharing the code publicly or with their team. Then, there are outer development loop, which starts as soon as your code is pushed into version control (or you send a PR for the same). This is where you hand over most of the responsibilities to automated systems and CI/CD pipelines to do the job for you as per a typical GitOps workflow.
Development workflows will vary from organization to organization. For example, the figure below demonstrates the hivenetes/bots application workflow. The goal is to take inspiration from it and build your optimized developer workflows.
If we had to break down the flow, it would be
Inner development loop
Outer development loop