I know it’s been a while. This afternoon is the first in weeks that I’m had a few moments to reset my brain.
I’m neck deep in an Azure POC for my business unit. I’m demoing a scalable, globally deployed microservice enabled system. (I know too many buzzwords). Think Excel client with a full cloud backend with everything that implies and it’s available worldwide.
There has been a lot to learn for this POC. But it’s all coming together. Along the way we did a full DevOps pipeline on VisualStudio.com. I learned some of the ins and out for release a service fabric app to secured cluster; added more PowerShell skills to my toolbox; extended my understanding of Docker and Azure Function; etc, etc, etc.
It’s been an interesting quarter.
In the last year or so, I’ve had it hammered into my head that the business demands value. In the past, I was more of a “pure” programmer. Not that I have changed all that much. What’s different is an understanding of the bigger picture of “business.”
My first question has become, “What’s the business value?” I still live and breathe research. I spend most of my week writing prototypes and developing infrastructure to accelerate the development of my teammates. But when it’s time for the rubber to meet the road, there has to be business value.
For instance, why move our code to Visual Studio Online? Well, to be blunt, the provided corporate infrastructure is shit. The system is down about as much as it is up. My team simply can’t function that way. So we moved off the corporate source control system to our own self-hosted TFS setup. That too became a problem when the composition of our team changed. No one wanted to maintain the servers. To put it another way, is there any business value in a development team supporting it’s own source control system? No. That’s not what we are paid to do. We’re paid to solve business problems and add value to the org. Futzing around with VM, TFS configurations, and a constant stream of patches and updates is far beyond our mission statement.
I did the basic research on moving up to VSTS. One of my teammates took it farther and we proved it out. Issues like permissions, internal deployments, click-once packaging, Coded UI testing and cloud deployments were resolved. Everything worked and now we’re migrating all of our code into the cloud.
Moving our source control into Azure enabled me to push releases into Service Fabric without any effort. Yes, I am running a direct pipeline into my POC SF instance. I check in my code. It is built. Upon a successful build it is released into my target SF instance automatically. Full CI/CD/CR integration. Now my POC users can get their code updates in minutes. Commit -> Build -> Release. And my manual management of the process ends when I check in my code because that is how I set it up. For our production systems, we enabled the approval process for all steps. Now my BAs fully own the final approval and release of an app into production. We also have gated check-ins to keep knuckle heads from checking in shit code. It really does work. And my life is much easier for the effort expended.
I should get back to whatever is next in my queue. I think I’m supposed to review some UX training videos. Yawn. Don’t forget I have a BA in Anthropology. I know a thing or two about studying group dynamics and interactions models. But still, there is always more to learn.