4Developers 2020, Brian Benz, DevOps: Improving applications with DevOps

### Overview In this presentation from 4Developers 2020, Brian Benz discusses how organizations can improve software delivery and operational efficiency through DevOps. Highlighting the balance of people, processes, and products, the talk emphasizes cross-functional collaboration and automated tooling to enable continuous value delivery to end users. Using statistics from the 2019 Accelerate State of DevOps report, Benz illustrates the impact of high-performing DevOps practices on deployment frequency, lead times, failure rates, and incident recovery.

### Key Topics - **DevOps Core Definition**: Grounded in Donovan Brown's definition—the union of people, process, and products to enable continuous delivery of value to end users. - **Software Development Life Cycle (SDLC)**: Structuring team collaboration across planning, developing, testing, releasing, and monitoring stages with distinct roles for dev and ops teams. - **DevOps Research and Assessment (DORA) Metrics**: Reviewing 2019 Accelerate findings, including 208x more frequent deployments, 106x faster commit-to-delivery time, 7x lower change failure rates, and 2604x faster incident recovery. - **Tooling Ecosystem**: Examining Azure DevOps components (Boards, Repos, Pipelines, Artifacts, Test Plans) alongside GitHub, GitHub Actions, and third-party integrations like Jenkins. - **Deployment Strategies & Governance**: Utilizing deployment slots for canary releases, sticky sessions for A/B testing, and automated pre- and post-deployment approval gates.

### Technical Details Benz demonstrates a live end-to-end continuous integration and deployment workflow for a retail sample site named Tailwind Traders: - **Source Control and Triggering**: Editing pricing code locally in Visual Studio Code, staging, committing, and syncing changes to GitHub. The push initiates automated workflows. - **Build Pipeline Execution**: Using an Azure Pipelines YAML file targeting Microsoft-hosted build agents (such as Windows Server 2019 with Visual Studio). The build tasks invoke the NuGet tool installer, fetch Node.js, run Visual Studio build tasks, perform tests, and copy deliverables to an artifact staging directory. - **Intentional Build Failures**: In a parallel GitHub Actions workflow, Benz demonstrates failing a deployment run deliberately by mismatching Node version requirements (Node 14 vs. Node 10 required by the environment). - **Deployment Pipeline Architecture**: Staging moves artifacts across a three-stage release pipeline (Staging, Canary deployment slot, and Production). Benz explains how canary deployment slots support traffic splitting for A/B testing. Release progression is safeguarded by pre-deployment conditions, manual approvals, and automated policy gates (e.g., Azure Policy compliance, Azure Functions, or external REST API validation). - **Preserving Existing Investments**: Highlighting Jenkins integration architectures, where Jenkins handles initial builds and testing (leveraging Azure Container Instances or AWS EC2) before handing off artifacts to Azure DevOps for enterprise governance and release deployment.

### Takeaways DevOps success relies heavily on cultural alignment, shared responsibilities, and clear role definitions before tooling can be effective. Modern cloud toolchains enable teams to safeguard production through gated releases, deployment slots, and canary rollouts while maintaining high deployment velocity.