July 2023 Summaries

The blog post is part of a series on Infrastructure as Code (IaC) best practices, focusing specifically on the fictional company Zephyr Archaeotech Emporium's use of Pulumi to manage their online retail store's infrastructure. It highlights how Zephyr evolved from using simple Pulumi stacks to adopting a more complex, multi-project setup, which includes a newly added "zephyr-data" project for managing data infrastructure with Amazon Aurora. To handle the increased complexity and potential deployment errors, Zephyr's team employed Pulumi's Automation API, which allows for programmatic control of Pulumi operations using various programming languages, simplifying the orchestration of multiple stacks. The post emphasizes separating configuration from code for easier distribution and using local source to enable independent management by separate teams, while inline source may be preferable when tight coupling is necessary. The Automation API program is written in Golang and enables efficient stack operations like updates and destruction, simplifying the deployment workflow for Zephyr's team.

Pulumi has introduced a feature in the Pulumi Cloud console that allows users to restore previously deleted stacks, addressing customer demands for recovering stacks due to accidental or intentional deletions. This new functionality enables users to preserve the activity history of a stack while removing its resources, contrasting with existing commands like `pulumi destroy` and `pulumi stack rm`, which either delete resources but keep the state file or remove the state file entirely. The new restoration feature, accessible through the Stacks page via a three-dot menu, allows organization admins to recover up to the last 25 deleted stacks, with older stacks requiring support assistance. This enhancement leverages Pulumi Cloud's capability to maintain state file versions, making the recovery process user-friendly and efficient, while encouraging users to provide feedback through the Pulumi Cloud Requests repository.

Pulumi has been deeply integrated with Kubernetes since its inception, offering comprehensive support across all versions of the Kubernetes API, and now Pulumi is releasing Kubernetes 4.0, which includes enhancements like default Server-Side Apply, upserting on all resources, and new Patch resources. The update simplifies access to outputs in various SDKs, supports Kubernetes clusters version 1.13 or newer, and enhances the diff process to detect resource drift more efficiently. These improvements make Pulumi's platform more native to Kubernetes, allowing for flexible resource management and integration with Pulumi's infrastructure-as-code model in popular programming languages. The release aims to streamline the management of Kubernetes resources even when they are modified by external systems, and Pulumi's new features facilitate safer multi-controller management and more precise updates to shared resources. Users can migrate to Kubernetes 4.0, benefit from improved performance for larger stacks, and take advantage of the free on-demand webinar offered by Pulumi and the CNCF to explore these advancements and build internal developer platforms.