March 2020 Summaries

Kubernetes provides scalability and reliability for container-based applications, and when combined with Google Kubernetes Engine (GKE), it removes the need for users to manually install or operate clusters. Running Kubernetes within a Google Cloud Platform (GCP) environment offers numerous integrations and benefits, particularly when using Codefresh, which automates code building, testing, and deployment processes. The guide highlights a step-by-step approach to deploying applications to Kubernetes by first creating a cluster in GKE through the GCP Console or Google Cloud CLI, then connecting Codefresh to GKE, and finally adding clusters to Codefresh for deployment. Codefresh simplifies the deployment of Docker images to Kubernetes clusters and provides an Environment Dashboard for monitoring deployments. Users who sign up through the tutorial also receive $500 in GCP credits, encouraging exploration of these tools within their DevOps pipelines.

Codefresh has introduced several significant enhancements to its platform in conjunction with the removal of its Docker registry, aiming to improve traceability and flexibility in pipeline workflows and artifact storage. Key features include the new Image API and an enhanced pipeline build step, designed to provide full traceability from source code commits to production deployment, thus allowing users to track the entire software lifecycle within a single platform. Codefresh supports multiple cloud environments and external Docker registries, providing a centralized dashboard that displays images from all connected registries, which eliminates the need for users to check multiple platforms. Additionally, Codefresh has updated its auto-push behavior, allowing users to select which registries are used for different stages of development, offering greater flexibility and control over deployment workflows. These enhancements are meant to streamline operations, especially for organizations utilizing Helm alongside Docker and Kubernetes, ensuring a comprehensive view and management of the entire release process.

Codefresh is set to deprecate its built-in container registry in favor of focusing on its core CI/CD platform features, with the registry becoming read-only on July 1, 2020, and completely disabled by July 15, 2020. This decision aligns with the increasing adoption of external registries by customers and the company's aim to prioritize best-in-class integrations over maintaining its own registry. Codefresh assures that the removal will not impact pipeline performance, as existing features like Docker caching will still function with external registries. Customers are encouraged to transition to external registries, with recommended options including Azure’s ACR, Google’s GCR, and AWS ECR, among others. Codefresh is providing detailed documentation and dedicated support to assist with the migration process, emphasizing that the change allows them to allocate resources towards new features and improvements.

The text discusses the evolution of Java application deployment from using single deployment units like WARs, JARs, and EARs to adopting microservices architecture, facilitated by Docker. It explains how Docker simplifies the process of upgrading services by allowing developers to create new Docker images with updated deployment units instead of redeploying entire applications. The author reviews three methods for building Docker images for Java applications: package-only builds, normal Docker builds, and multi-stage builds, highlighting multi-stage builds as the most efficient approach. Multi-stage builds create smaller images by only including necessary components, which is beneficial for cloud storage. The text also includes a practical guide with commands and a link to a GitHub repository for users to follow along, promising future discussions on handling caching issues in Java builds.