Interview with Lori Lamkin on Microsoft’s DevOps Journey and Practices

This article is an interview with Lori Lamkin, Program Management Director for Microsoft Visual Studio Cloud Services, who leads a team of project managers driving online developer experiences, cloud‑connected applications, collaboration, community engagement, and modern development practices.

In 2017, Microsoft’s VSTS DevOps journey was already underway, aiming to reduce cycle time, adopt DevOps, and keep pace with rapid business changes.

1. Change the Rhythm – The first step was to accelerate release cadence, initially introducing a "stabilization" phase which proved counter‑productive; abandoning it encouraged teams to avoid accumulating technical debt.

2. Team Autonomy and Organizational Alignment – Empower teams to own their backlog and planning, fostering transparent communication and KPI discussions directly with product teams.

3. Continuous Delivery – Commit to delivering features on a fixed three‑week iteration schedule, continuously listening and learning.

4. Trunk‑Based Development – Move away from long‑lived branches to low‑cost Git branching and develop on the mainline to reduce integration overhead.

5. Dogfooding – Use the product internally, improving productivity and product quality.

6. Continuous Deployment – Increase deployment frequency with smaller batches to reduce deployment debt and improve tool reliability.

7. Record Everything – Shift from blind debugging to comprehensive logging and telemetry.

8. Continuous Testing – Make developers responsible for testing, eliminating hand‑offs that cause delays and waste.

9. Trustworthy Tests – Provide fast, reliable feedback on pull requests and CI runs.

10. Portable Tests – Ensure tests run quickly (e.g., 55,000 tests in six minutes) and provide clear signal on code changes.

11. Feature Flags in the Cloud – Use feature toggles to enable safe, incremental releases without accruing debt.

12. Resilient Architecture – Prevent cascading failures and mitigate noisy neighbor issues.

13. Security – Foster a security‑first culture through red‑team/blue‑team exercises and realistic vulnerability simulations.

14. Engineer Ownership of Production – Shift responsibility for production stability to development teams, reducing reliance on separate ops teams.

15. Production Telemetry – Build actionable alerting systems and self‑healing mechanisms to handle incidents efficiently.

16. Data‑Driven Business – Embrace metrics and telemetry to drive decisions and improve service health.

17. Meaningful Metrics – Design metrics that reflect true performance, focusing on speed and quality rather than vanity numbers.

18. Telemetry‑Verified Completion – Use telemetry to confirm feature adoption and guide backlog priorities.

19. Ongoing Journey – Recognize that DevOps transformation is a continuous, evolving process that reshapes development, testing, and operational responsibilities.