Where are we in Donovan’s [from 0 to DevOps] world?
If you have not watched from 0 to DevOps, by Donovan Brown, or read Applying DevOps to a Software Development Project, you should pause this post, peruse the material, and return.
Find the supporting presentation here.
OK, now that you’ve met 
, we continue with our how has the Ranger community evolved series. Although we have applied learnings on becoming more Agile for years, we created an operational bottleneck with our wave of extensions for Visual Studio Team Services. We realized that we had to become more effective with our builds, releases, monitoring, and the ability to react to user feedback. As recommended by Donovan Brown, we “rubbed a little DevOps” on our people, process, and all of our extension projects.
Some of our key initiatives we covered in other publications:
- Instrumentation of all extensions – Application Insights tokenization for Visual Studio Team Services extensions
- Visualization of day-to-day operations and backlog – The secret sauce to managing our backlog and TODOs
- Automated continuous integration (CI) and continuous deployment (CD) pipelines – Applying DevOps to a Software Development Project
- Getting to grips with lead and cycle times – Understanding Lead and Cycle Time for our part-time community
Now it’s time to reflect. We’ll review how the DevOps assessment saw us and how we view ourselves.
Mapping 0 to DevOps – our perspective
When we got together in a recent “bosberaad”, we mapped Donovan’s crisp definition of DevOps to our ecosystem as shown.
We view ourselves as maturing in terms of people, value, products, and continuous delivery. For process we’ll innovate continuously for a while ever. In the context of DevOps we view ourselves as “healthier, not sick or healthy”.
Mapping 0 to DevOps – DevOps Assessment
Next we evaluated ourselves using the DevOps Assessment a number of times. As shown, our assessments varied only for Production and Flow, tying back to the process mentioned previously.
Although the assessment results were really promising, we’re not satisfied. We believe that we have to innovate and mature in all the shown categories, and have to continue doing so. The assessment is a great way to get a perspective on how you’re doing with DevOps, but more importantly an opportunity to have a focused “team” discussion.
Based on the assessment, we need to focus on:
- Reduce lead + cycle times
- Reduce mean time to repair (MTTR)
- Implement A|B testing
Mapping 0 to DevOps – Visualising our progress
We could have picked any platform, any tools, and any language, as outlined in Best of Both Worlds. As “Visual Studio ALM Rangers” it comes as no surprise that we picked Team Services as a unified solution for planning, continuous integration (build), continuous delivery, and monitoring. We plug and play with a variation of tools, to implement the most efficient pipeline for each project, some using Visual Studio, some Visual Studio Code, as well as a variety of other products.
Gone are the days were teams notified the build master to fire up the build. We’re in a new world where a check-in triggers a platter of automated tasks, stopping only for approvals to deploy releases to evaluation, testing, and finally production environments.
Using telemetry and pro-active detection, we’re able to react to issues quicker, even though our teams are part-time, geographically distributed, and often focused on new projects. We can also analyse usage and identify interesting patterns, as shown in this snippet below.
The world DevOps is your Oyster!
OK, where to from here?
DevOps is about a mindset of reflection and continuous innovation, with an objective to ship solutions (value) more frequently. Our team is busy researching how to use Azure DevTest Labs to simulate and test on-prem environments, and WhiteSource to detect and validate open source components as part of our pipelines.
Watch the space!