How we work and lessons we’ve learned building Team Services and TFS


Here are two recent presentations that discuss the evolution of our team from an on-premises software team to a DevOps services team.

The first one is a presentation by Matt Manela and Jose Rady Allende at Microsoft Ignite last week. Matt is the engineering manager and Jose the PM for one of the feature teams building the Agile features in Team Services and TFS. They provide a first hand view of how the way we work has changed as well as what’s worked well for their own team. Not only does the talk cover how we worked before DevOps, how the team is organized, and how we do planning, it also covers our test philosophy, the engineering metrics we track, and a team organization approach they’ve found to work well for their feature team. Even if you seen a past presentation like the one Lori and Brian gave at Build, there’s new material in it. You can find it at Microsoft Ignite On-Demand and on YouTube.

The second is a presentation that I gave two weeks ago at the ITARC Southeast 2016 conference. In it I talk about

  • the evolution of the service, including deployment including online database upgrades,
  • using feature flags to control exposure,
  • circuit breakers to prevent cascading failures,
  • and where we are headed in trying to detect indications of problems within the large volume of telemetry we collect.

This talk was also the basis for a recent blog post on feature flags. You can find the presentation on YouTube (slide deck).

The two talks cover different material with almost no overlap and together provide a lot of insight into how we work and the lessons that we’ve learned along the way.


Follow me at


Discussion is closed.

Feedback usabilla icon