January 2022 Summaries

Pulumi has introduced a range of new features and improvements in its latest release, enhancing various aspects of its cloud infrastructure management capabilities. The update includes a new Pulumi Command package that facilitates running scripts locally or remotely as part of the resource lifecycle, now available in all Pulumi languages. Enhanced support for Kubernetes CustomResourceDefinitions (CRDs) and significant improvements to Helm Release, including importing existing Helm releases and using YAML files for Helm values, have also been added. Additionally, the Pulumi CLI now supports native ECMAScript modules (ESM) for Node.js SDK, allowing top-level await in Pulumi programs. Pulumi packages can now host plugins in third-party locations, and state locking is enabled by default for self-managed backends. Audit logging has been expanded to include third-party secrets managers, ensuring comprehensive event logging for users not using Pulumi's secrets provider.

Pulumi Service, the state-management backend of the Pulumi CLI, offers a range of features for managing infrastructure, including stack tags for organization, linking updates to CI/CD pipelines, customizing update messages, viewing resource changes, visualizing stack resources, and drilling into resource properties. Users can manage stack tags through both the CLI and the Pulumi Console, improving project organization and resource management. The service supports linking stack updates to CI/CD jobs and source commits for better traceability and allows customization of update messages to provide context for infrastructure changes. Different views, such as Diff and Diagnostic, offer insights into resource modifications and update logs, while the Stack Resource Visualization feature aids in understanding resource relationships through an interactive graph. Users can explore resource properties in detail and access cloud provider consoles directly from Pulumi, with ongoing updates and community support enhancing the service's capabilities.

Kay Plößer explores setting up observability for Pulumi deployments using Honeycomb, emphasizing the integration of these tools to monitor infrastructure effectively. Pulumi, a cloud-provider independent infrastructure as code (IaC) framework, allows defining infrastructure in TypeScript and supports various providers. By leveraging Pulumi's Automation API, users can simulate CLI actions programmatically, capturing deployment events like start times, resource deployments, and errors, and sending this data to Honeycomb for detailed trace analysis. This integration helps in understanding deployment behaviors, resource impact on deployment time, and diagnosing errors, enhancing the visibility and reliability of infrastructure operations. The article provides a step-by-step guide to configure Honeycomb with Pulumi, using Honeycomb's Libhoney client for sending deployment data, thereby facilitating real-time insights into the deployment process.

Originating from a hackathon project, this blog post explores potential solutions to the challenge of resolving circular dependencies in Pulumi programs, a common issue when deploying modern web applications. The text provides a detailed examination of four conceptual solutions: using a "State" object to manage values across deployments, employing "Refs" to integrate dependency tracking into Pulumi's graph, creating "patch" resources to model updates explicitly, and allowing resources to update their goal state dynamically. Each approach is assessed for its feasibility and drawbacks, such as deployment complexity, potential for endless update loops, and increased provider size. While none of the solutions are currently implemented in Pulumi SDKs, the exploration offers valuable insights into managing dependencies and invites further contributions from the community, highlighting the project's open-source nature and ongoing development.

The text discusses the integration of AWS services to create a scalable and maintainable application architecture using Amazon API Gateway, EventBridge, and AWS Lambda, with a focus on overcoming the limitations of tightly coupling API Gateway with Lambda for handling HTTP requests. It highlights how to use EventBridge as a serverless event bus to decouple and manage event-driven processes, allowing multiple independent actions in response to a single API call. The guide provides a step-by-step approach to building this architecture using Pulumi, detailing the setup of API Gateway, EventBridge, and Lambda functions, and emphasizes the use of AWS_PROXY integration type for efficient communication between these services. The process involves creating an event bus and event rule to handle notifications and employing Pulumi for infrastructure management to facilitate the development and deployment of serverless applications. The text concludes with instructions on verifying the setup and suggests further enhancements and integrations to explore, such as adding more Lambda functions or targeting additional AWS services.

Josh Graham, a seasoned SaaS architect from Atlassian, explored using Pulumi and LocalStack to create a serverless application, highlighting the benefits and challenges of using these tools for infrastructure-as-code. In his endeavor, Graham utilized AWS services like DynamoDB and Lambda, leveraging Pulumi's TypeScript-based infrastructure management to streamline the process. He emphasized separating infrastructure code from application code to enhance modularity and testing, and he addressed the complexities of handling provision-time and runtime values in Pulumi. Additionally, Graham shared insights into using LocalStack for local AWS service emulation, which aids in faster feedback and cost reduction, while also noting some technical hurdles encountered, such as ensuring AWS credentials are available in LocalStack's Docker containers. Through his project, Graham demonstrated the potential for organized, test-driven development in serverless architectures, providing a template for others to follow and highlighting the evolving capabilities of Pulumi when used with LocalStack.

Pulumi's December 2021 hackathon brought together teams from across the company for a week-long event focused on improving the Pulumi ecosystem, with both internal and open-source projects. The preparation involved early discussions on Slack to allow participants to choose projects and engage in architectural planning, fostering excitement and early collaboration across different departments. As a remote workplace, Pulumi adapted to the challenges of virtual collaboration by allowing teams to select their preferred communication methods, including Zoom standups, Slack interactions, and pair programming. The hackathon facilitated effective teamwork across time zones, allowing participants to learn from each other, brainstorm solutions, and connect without the typical drawbacks of virtual interactions. The event concluded with a demo where teams shared strategies, insights, and project outcomes, highlighting the success of pairing team members based on location and work preferences. Pulumi is also extending an invitation for new job applications to join their innovative working environment.

Pulumi Registry offers a solution to common AWS infrastructure challenges by providing higher-order components that simplify configuring, deploying, and managing AWS resources, addressing issues such as AWS Console Wizard Fever, YAML/JSON CloudFormation Template Ceiling, and AWS Lock-in. The registry abstracts the complexity of AWS services, making it easier to implement best practices in resource management without the steep learning curve associated with traditional Infrastructure as Code methods. By allowing developers to use programming languages instead of YAML/JSON templates, Pulumi enhances flexibility, reduces overhead, and provides infrastructure review tools to prevent deployment errors. It also facilitates the integration of non-AWS tools, enabling service providers to compete with AWS's native services. Pulumi's Quick Start components, like those for VPC, Aurora Postgres, and Redshift, are designed to be user-friendly for both novice and experienced AWS developers, offering a streamlined and reliable approach to deploying intricate AWS architectures.