CI/CD with Robert Erez

2026-06-03 · Source: The Pragmatic Engineer · Field: Technology & Digital — Software Development & Engineering, Cloud Computing & IT Infrastructure, Artificial Intelligence & Machine Learning · Depth: Advanced, extended

Summary

Robert Erez of Octopus Deploy discusses the evolution of CI/CD, highlighting early continuous delivery practices at Skype, including using New Zealand for canary deployments. He defines the stages from YOLO to continuous deployment, noting that full continuous deployment isn't suitable for all regulated industries. Erez explains Kubernetes' rise as a cross-platform container orchestration winner, used both cloud-native and on-premise, and its declarative state management. He clarifies GitOps' four pillars (declarative, versioned/immutable, pull, continuous reconciliation), emphasizing that not everything, especially secrets, belongs in Git. The discussion also covers the emergence of platform teams for standardizing deployment workflows and AI's anticipated impact on CI/CD, predicting higher code velocity and increased reliance on progressive delivery techniques like feature toggles for risk mitigation. Octopus Deploy's SaaS and on-premise operational challenges are also explored.

Key takeaway

For DevOps Engineers managing complex deployments, prioritize progressive delivery strategies like feature toggles to decouple feature releases from code deployments. This allows for rapid, controlled rollouts and immediate rollbacks of individual features, significantly reducing risk compared to full application rollbacks, especially with stateful systems. Additionally, embrace GitOps principles for declarative infrastructure management, but avoid dogmatically putting all configurations, particularly secrets, into Git.

Key insights

Effective software delivery balances automation, controlled progressive releases, and adaptable infrastructure, prioritizing practical outcomes over dogmatic approaches.

Principles

• Continuous delivery progresses from YOLO to full automation.
• Declarative state management simplifies infrastructure orchestration.
• Decouple feature release from code deployment via toggles.

Method

Progressive delivery de-risks releases using techniques like canary deployments, blue/green strategies, and feature toggles to gradually expose changes and manage risk.

In practice

• Begin progressive delivery by implementing a single feature toggle.
• Utilize ephemeral environments for pre-merge branch testing.
• Prioritize rolling forward with hotfixes over complex rollbacks.

Topics

• Continuous Delivery
• Progressive Delivery
• GitOps
• Kubernetes
• Feature Toggles
• Platform Engineering

Best for: DevOps Engineer, Software Engineer, MLOps Engineer

Related on AIssential

• Must- Know Deployment Strategies: From Big-Bang to Progressive Delivery — Deployment strategies primarily aim to reduce risk and control user exposure during code releases.
• GitOps for Low-Code: Engineering a Production-Grade CI/CD Pipeline for n8n on Kubernetes — Low-code platforms such as n8n are increasingly vital for automating workflows and integrating APIs, but their ad-hoc deployment methods prove unsustainable for mission-critical production environments.
• A Beginner’s Guide to Kubernetes — Kubernetes maintains desired states through a system of continuous "promises" rather than sequential instructions.
• Can We Tackle Security & Compliance in Our Deployment Pipeline? — Continuous delivery is essential for automating and enforcing robust security and compliance, making systems inherently safer and more auditable.
• From GitOps to AgenticOps: Giving Kubernetes a Brain — AgenticOps layers AI reasoning onto GitOps and Kubernetes for adaptive, goal-driven operational intelligence.
• How I Built a Scalable Feature Pipeline for Machine Learning — Reliable feature delivery, not feature engineering, is the core challenge in production ML systems.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Pragmatic Engineer.