August 2019 Summaries

Tilt.dev, developed by Windmill Engineering, is a tool designed to facilitate the local continuous development and deployment of microservice applications, focusing on the "inner-loop" of development workflow by deploying applications to local Kubernetes environments like minikube, docker-for-desktop, or microk8s. Similar to other tools such as Draft, Skaffold, and Garden, Tilt offers features like automatic deployments, a command line, and a web UI for monitoring application deployment and logs, aiming to streamline development without needing multiple terminal windows. It supports more than just Dockerfiles, allowing developers to reuse Kubernetes manifests and integrate with Helm and Kustomize, thus enhancing the efficiency and speed of developing and deploying complex microservices with interdependent Docker images. Tilt's distinctive features include a live dashboard for simultaneous microservice monitoring and a Python-like DSL for configuration files, making it particularly suitable for Kubernetes-specific applications. In contrast to Garden, which provides a project-level approach with a dependency graph and supports in-cluster building, Tilt focuses on watching for changes across services and performs live updates directly in a local workstation's environment, streamlining the build, push, and deploy processes in real time.

Debugging microservices in Kubernetes presents significant challenges due to the difficulty of replicating microservice environments locally and the unstable access to pods when using traditional debugging tools. While Kubernetes recommends debugging within a remote shell on a running container, this method is often cumbersome and unfamiliar to many developers who prefer local debugging. Although there are solutions available for simulating microservice environments on local machines, none provide a perfect solution. Rookout offers a promising alternative by enabling rapid debugging across development, staging, and production environments without the need for additional code, restarts, or redeployment. It supports multiple programming languages and allows developers to set non-breaking breakpoints to obtain data instantly, ensuring that applications continue to function uninterrupted. Despite these options, the complexity of debugging in Kubernetes remains, prompting a consideration of remote debugging methodologies.

In the current technology landscape, where clients demand constant and seamless updates, continuous delivery (CD) plays a crucial role in optimizing the time to market for modern applications. Continuous Integration (CI) and CD practices, as defined by experts, involve integrating code into a shared repository multiple times a day and ensuring every system change is releasable with the push of a button, respectively. These practices help in detecting problems early and making the release process routine and predictable. A well-constructed deployment pipeline, which includes stages like unit and integration testing, security scans, and deployment, ensures consistent and repeatable results across various environments, thus reducing errors and inefficiencies. The temptation to build in-house CI/CD tools can lead to complex, unmanageable systems, so it is advisable to use existing comprehensive solutions like Codefresh, which offer flexibility and visibility, making the deployment process 'boringly' reliable and efficient.