May 11, 2017 // By Magenic
Magenic’s leadership showed that small changes can make a big impact
Through automation, the time to create a set of builds went from a six-hour developer ordeal to a twenty-minute automated workflow.
An international financial services firm was attempting to modernize a suite of enterprise applications, but one was lagging behind. The company was struggling with their development model of off-shoring the work while keeping the management stateside. The offshore team could keep up with biweekly deliveries to development and test environments, but the increased cadence of bug-fixes was proving impossible to handle. As the release window was coming closer, a few manual processes were taking attention away from the team leads at a critical time. The client tasked Magenic with improving the speed and consistency of delivery.
Magenic stepped in and applied our best DevOps thinking. We asked if the developers were aligned with the goals of the business. We looked carefully at automation and tooling that could stabilize and optimize the output of the team. We researched the metrics and measurements that could give the business visibility into the group’s progress, and we connected them with other development teams that had worked through some of the same problems. Ultimately, we found that the bottleneck was a manual build process that relied entirely too much on inconsistently configured developer machines. We liberally applied tools like Team Foundation Server, PowerShell, and WiX to create a pipeline that didn’t involve nearly as much intervention.
Stateside managers and stakeholders were happy to see successful builds happening on every check-in. They learned that the perceived instability was coming not from the people or their code, but the process. Once testing was automated, the business was able to provide feedback and request fixes multiple times per day. Not every developer group needed to adopt new and unfamiliar tools to get the benefits of DevOps thinking. Only by focusing on the bottleneck were we able to increase the overall throughput of work in the system.