March 2026 Summaries

In the second installment of the "Practical Platform Engineering in 5 Lunches" series, readers learn how to create a simple Kubernetes deployment project using Octopus Deploy and its AI Assistant. The post guides users through setting up a mock Kubernetes cluster, creating necessary resources like a Kubernetes project, token account, and Docker Hub feed, and configuring Platform Hub with a mock Git repository. Additionally, it explains how to publish and share process templates within Octopus, allowing for streamlined updates and distribution of architectural decisions across projects. This hands-on experimentation with mock environments enables users to explore Octopus's features without access to real infrastructure. The series aims to educate on building Internal Developer Platforms, with this post advancing the process by demonstrating practical applications of Octopus Deploy's capabilities.

Matthew Casperson's blog post series "Practical Platform Engineering in 5 Lunches" guides readers through developing an Internal Developer Platform (IDP) using Octopus Deploy. Designed to be completed within lunch breaks, each post takes about 30 minutes, requiring only a free trial of Octopus and no additional platforms. The introductory post establishes the basic prerequisites and concepts, including definitions of DevOps, Platform Engineering, and IDPs. Casperson argues DevOps has become a broad, unfalsifiable term, while Platform Engineering is about scaling architectural decisions to streamline CI/CD processes. Octopus supports Platform Engineering through a Platform Hub, which leverages Git workflows for maintaining and distributing shared processes and policies. By the end of the post, readers have set up a trial Octopus instance, installed the Octopus AI Assistant Chrome extension, and gained foundational knowledge for the series, with future posts delving into Kubernetes deployment projects.

Since its Early Access release in 2025, the integration of Argo CD into Octopus has seen incremental feature additions, with the latest being step verification, which ensures the health of Argo CD applications before completing an Octopus step. This feature provides multiple options for step progression, including direct commits, pull request merges, and the health status of Argo CD applications, enhancing the accuracy of the Octopus dashboard in reflecting deployment status. The introduction of "Git drift" scenarios highlights various synchronization states between Octopus and Argo CD, emphasizing the importance of verifying whether changes have been applied to the cluster. Octopus tracks intended changes by recording JSON patches and file hashes, allowing it to determine if its updates have been overwritten. While the step verification feature is readily available to customers from version 2026.1, the pull request merged verification will be available with version 2026.2. The integration with Argo CD compliments the tool's GitOps capabilities by adding functionalities like environment promotions and orchestration, providing a comprehensive Continuous Delivery platform without the need for custom automation.

Steve Fenton's article explores the contrasting approaches of FinCost and FinBoss organizations regarding developer experience and resource investment. FinCost entities prioritize visible cost-cutting measures, often leading to underinvestment in essential tools and training, which can result in long-term inefficiencies and higher costs. In contrast, FinBoss organizations focus on aligning finance and leadership to maximize return on investment by providing developers with high-quality equipment and opportunities for skill development, thereby enhancing productivity and long-term success. Drawing on insights from DORA's research, Fenton emphasizes the importance of transformational leadership, a supportive organizational culture, and strategic investment in tools to foster an environment where developers can perform optimally. The article urges organizations to evaluate their current practices and consider the tangible benefits of adopting a FinBoss approach to developer experience.

Octopus has introduced Starter Policies to simplify the use of Rego, the language of the Open Policy Agent (OPA), for defining and enforcing compliance in software deployments, available through the Enterprise tier's Platform Hub. The guided setup process allows users to select from a library of common compliance use cases, generating boilerplate code that can be customized to fit specific team needs. These policies ensure governance, such as requiring manual approval for production deployments, and provide inline comments for better understanding of the Rego logic, helping teams to learn by doing. The Platform Hub serves as the centralized location for managing software delivery, insights, and governance, with all policy evaluations and outcomes being recorded in an audit log. Violation actions can be configured to either block non-compliant actions or allow them with a warning, depending on the team's compliance requirements.

Inside Platform Engineering, a conversation series hosted by Matt Allford, features Piotr Szwed discussing the evolving landscape of platform engineering, highlighting a shift from traditional automation to autonomous systems that are API-first and AI-native. Piotr emphasizes the limitations of current tools like Terraform in keeping up with the rapid expansion of cloud services and advocates for the adoption of systems that manage themselves, akin to the Kubernetes operator pattern. He notes that reference architectures from major players like Amazon, Cisco, SAP, and Google now provide modular, API-first frameworks that can be customized to fit specific organizational needs, signaling a convergence in industry practices. The discussion underscores the importance of starting the transition to these modern platforms incrementally, leveraging mature ecosystems and active communities to navigate the change.

Steve Fenton emphasizes the importance of maintaining a deployable state for software to minimize risk and ensure smooth operations. He argues that unreleased code accumulates risk and that prioritizing deployability over feature development can prevent issues such as overheads, rework, and failures. A robust deployment pipeline, which includes automation of tests and dependency checks, is crucial for swiftly identifying issues and ensuring readiness for deployment. Fenton highlights the limitations of manual testing, advocating instead for automated testing as a foundation that supports sustainable software development. He underscores the value of treating automated tests and deployment processes as living documentation, which aids in knowledge transfer and reduces the risk of undocumented changes causing deployment failures. Continuous Delivery practices, while requiring initial effort, contribute to long-term sustainability and economic viability by reducing maintenance costs and defect density.

Octopus Deploy has enhanced its tag sets feature, which is crucial for tenanted deployments, by allowing tags to be applied to projects, environments, and runbooks in addition to tenants. The updated system supports three types of tag sets: MultiSelect, SingleSelect, and FreeText, each designed for different scenarios such as selecting multiple options, ensuring a single choice, or allowing custom text values. Project tags can be used to assign metadata like teams or technology stacks, making dashboard configuration more relevant, while environment tags help categorize environments by production status or cloud provider, aiding in policy assignments. Runbook tags allow categorization and event-based triggering of runbooks. The document advises on best practices for designing tag sets to streamline filtering and management and notes ongoing efforts to migrate existing target tags to the new tag set management, with potential future developments including deployment freezes and license usage tracking by tag.

In the detailed guide by Matthew Casperson, readers are walked through creating a mock Kubernetes deployment using Octopus Cloud and its AI Assistant. Users are encouraged to set up an Octopus Cloud account and install the Octopus AI Assistant Chrome extension, which simplifies the process of deploying to a mock Kubernetes cluster. The guide outlines how to construct a Kubernetes project titled "14. Kubernetes with Mock Target" by utilizing specific container images, enabling retries, and configuring a mock Kubernetes API server for testing purposes without needing access to a real cluster. It includes instructions for setting up a Kubernetes target and executing deployment steps within a Hosted Ubuntu worker pool. The mock server facilitates local interaction and testing, while the project also incorporates a Software Bill of Materials (SBOM) scan to identify security vulnerabilities, with the results updated daily through a dedicated security environment. Once the deployment is executed, the execution container and mock server are discarded, demonstrating a comprehensive approach to learning and experimentation in Kubernetes deployment.

Argo CD 3.3, now generally available, introduces significant enhancements to improve scalability, usability, and the efficacy of GitOps workflows. Key features include expanded integration with Microsoft Teams for notifications, improving migration flexibility for teams transitioning from Office 365 Connectors, and enhanced health checks for third-party operators like Grafana. The release also introduces an abstraction layer for the AppSet UI codebase, preparing it for future integrations without altering current functionalities. New actions for CloudNativePG and KEDA resources allow for operational tasks and scaling behavior adjustments directly from the Argo CD UI or CLI. The CLI sees improvements with more examples, project filters, and PowerShell completion support, enhancing its usability. Internal updates include performance enhancements for the ApplicationSet controller, reflecting a commitment to better automation, operational visibility, and developer experience in GitOps processes.

In this blog post by Matthew Casperson, readers are guided through the process of defining custom lifecycles in an Octopus project, building on previous work where a custom channel was added. Key prerequisites include an Octopus Cloud account and the Octopus AI Assistant Chrome extension, with the cloud-hosted version being the simplest to use due to minimal configuration needs. The post provides instructions for creating a script project titled "13. Script App with Lifecycle," which includes setting up an "Auto Deploy" lifecycle with phases for Development, Test, and Production environments, and establishing a default channel named "Application." This setup enables automatic deployment to the Development environment upon release creation. The post concludes by suggesting the next step of creating a mock Kubernetes deployment project, catering to DevOps and AI enthusiasts.

In this post by Matthew Casperson, readers are guided through the process of defining project channels using the Octopus AI Assistant in a DevOps context. The tutorial requires an Octopus Cloud account and the Octopus AI Assistant Chrome extension for ease of use, though the cloud-hosted version is recommended for its simplicity. The guide walks through creating a new script project named "12. Script App with Channel," where a custom lifecycle called "Hot Fix" is defined to deploy directly to the "Production" environment, bypassing earlier stages. This approach is particularly beneficial for deploying urgent fixes directly to production without adhering to the standard deployment pipeline, providing a streamlined solution for critical updates. The next step in the process involves defining a custom lifecycle for further customization and efficiency in deployment strategies.