June 2026 Summaries

Jubril Oyetunji's guide explores the implementation of progressive delivery strategies in Kubernetes using Argo Rollouts, a Kubernetes controller that enables blue/green and canary deployments, enhancing the default rolling update strategy. The guide details the installation and configuration of Argo Rollouts, which provides real traffic shaping through integrations with ingress controllers and service meshes, and allows for monitoring via metrics providers like Prometheus and Datadog. It outlines a step-by-step process for setting up a blue/green deployment, allowing the user to manage active and preview services and handle promotions and rollbacks efficiently. The guide emphasizes the need for orchestration tools like Octopus Deploy to manage higher-level deployment processes, such as environment promotion and audit trails, complementing the cluster-side mechanics managed by Argo Rollouts and Argo CD. Through this setup, users can achieve a more reliable and less stressful deployment process, ensuring minimal disruption and quick recovery from potential issues.

Paul Stovell's article challenges the myth that speed alone drives software development success, emphasizing that while rapid deployment and automation can enhance throughput, they must be coupled with a deep understanding of user needs and consistent feedback loops to achieve true product-market fit. Using the Bullseye Model as an analogy, Stovell argues that like an archer aiming for consistency rather than just speed, software teams should focus on developing reliable processes that allow them to adjust their course based on user feedback and changing conditions. The article highlights that while AI-assisted development can boost throughput, it necessitates refined deployment pipelines and governance to avoid negating potential gains. Ultimately, the goal is not just to accelerate the development process but to move strategically towards an ideal product that meets customer needs, using speed as a tool for frequent course correction rather than as a sole objective.

Argo CD 3.5 introduces native mutual TLS (mTLS) support, enhancing the security of internal communications between its components by embedding encryption and identity verification directly into the application. This advancement addresses the limitations of one-way TLS and aligns with zero-trust architecture principles, ensuring that both client and server authenticate each other before data exchange, which significantly reduces security risks such as unauthorized access. Prior to this update, operators relied on complex external solutions like service meshes or custom scripts for encryption, which added operational overhead and maintenance challenges. The new mTLS feature simplifies this by enabling easy setup through a Kubernetes Secret, allowing for both shared and per-component certificate configurations, thus catering to various security needs and compliance requirements such as SOC 2, HIPAA, and PCI-DSS. The update facilitates smoother transitions from existing architectures and enhances traceability and security without the need for third-party dependencies, making it a valuable enhancement for diverse environments, from small-scale deployments to large, regulated enterprises.

At Octopus Deploy, Matthew Allford modernized the build and release processes for microsites by integrating ephemeral environments into the Pull Request (PR) process, which allows changes to be reviewed in a real, working environment. Although these environments add value, they are not always necessary and can delay the build process or incur high costs, especially in busy repositories. To address this, Allford refactored the GitHub workflow to make the deployment of ephemeral environments optional by using PR labels, allowing developers to skip this step when appropriate without affecting essential checks or branch protection rules. This change, implemented as part of a shared workflow infrastructure, benefits all teams using the same process, demonstrating the advantages of a Platform Engineering mindset where improvements are shared and propagated across teams.

Kostis Kapelonis explores the shift to using external secret controllers for managing application secrets within the GitOps paradigm using Argo CD, particularly emphasizing the advantages of the Secret Store CSI driver. With the release of Argo CD v3.x, there is a clear recommendation to use external secret solutions instead of those tied directly to Argo CD's manifest generation process. The Secret Store CSI driver offers a significant benefit by directly mounting secrets as files on pods without using Kubernetes secrets, aligning with certain security policies that prefer avoiding Kubernetes secrets entirely. This approach allows for seamless secret rotation without requiring application redeployments or Argo CD sync operations. The post also includes a demonstration using HashiCorp Vault to manage secrets within a Kubernetes cluster, showcasing how the CSI driver can facilitate secure and efficient secret handling.

As of release 2026.2, self-hosted Octopus Server customers can now utilize GitHub Connections through the Octopus Deploy App for GitHub, eliminating the need for Personal Access Tokens (PATs) which posed several security and operational challenges. This integration employs OpenID Connect (OIDC) to facilitate a secure token exchange between Octopus and GitHub, ensuring permissions are managed at the app installation level rather than tied to individual user accounts, which mitigates risks associated with team changes and token expirations. For instances behind firewalls, Octopus now offers an Externally Hosted option for signing keys, allowing users to publish keys on platforms like S3 or Azure, ensuring GitHub can validate tokens without requiring direct internet access to the Octopus instance. While GitHub Connections are available for repositories on github.com, those using GitHub Enterprise Server must continue using PATs, although interest in expanding support can be registered.

The text discusses how Octopus Deploy integrates with Argo CD to manage Kubernetes deployments, emphasizing the challenges of rolling back deployments in practice. It highlights that while Argo CD excels at synchronizing a Kubernetes cluster with a Git repository, it lacks built-in release management and rollback capabilities. Octopus Deploy addresses this by creating versioned snapshots of deployments that include the deployment process, variables, and referenced image versions, allowing for controlled promotion through different environments. The integration enables Octopus to manage lifecycle processes while Argo CD continues to reconcile resources, providing a streamlined approach to deployment and rollback that minimizes manual interventions. The text also offers insights into setting up a demo project using Octopus and Argo CD, detailing steps such as creating a project, adding Argo CD steps, setting up a manifests repository, and configuring Git credentials. Additionally, it explores the importance of backward-compatible database schema changes to facilitate safe rollbacks and suggests that automating deployment lifecycles can make rollbacks more reliable and less burdensome.

In the discussion led by Steve Fenton, the complexities and challenges of database changes compared to application code deployments are explored, emphasizing the high stakes and potential pitfalls of schema alterations. Unlike application code, database rollbacks are more intricate due to data dependencies and the risk of data loss, which cannot be entirely mitigated by backups. The text highlights the importance of version control for database schemas and the choice between state-based and migration-based update approaches, each with its procedural nuances. The role of tooling and automation in managing database changes is underscored, with a recommendation to avoid manual changes to ensure consistency and reliability. Further, it stresses the significance of automated test data management to prevent data inconsistencies during testing and proposes the expand/contract pattern for database refactoring to decouple application and database deployments, facilitating smoother transitions and rollback capabilities. Overall, the narrative encourages adopting best practices and robust tooling to transform database deployments into a stress-free and celebratory part of software delivery.

The partnership between Octopus Deploy and GitHub represents a pivotal shift in addressing the complexities of modern software development, where AI tools like Copilot are increasingly contributing code but also introducing new challenges in management and governance. By integrating Octopus Intelligence into GitHub's Agent HQ, the collaboration aims to streamline the software delivery process, enabling teams to manage deployment, troubleshoot issues, and maintain security more efficiently without the need for constant context switching. This integration allows developers to access and control coding agents from various AI providers directly within GitHub, facilitating a more cohesive workflow. The strategic alliance leverages innovations such as Copilot Extensions and Custom Agents to create an intelligent, AI-enabled software delivery platform that connects development, deployment, operations, and governance. Ultimately, this partnership is designed to foster a more connected software delivery experience, reducing friction and enhancing the speed and security of enterprise CI/CD landscapes, with Octopus orchestrating complex deployments across diverse infrastructures.