The State of Kubernetes in 2019 — Big Data Analytics-driven Market Research
As an industry analyst, I relish talking to the ‘guys in the trenches’ that see the brownfield side of the shiny vendor PowerPoint. Squeezing a commercial or open source software solution into your own use case, process requirements, and data center is a non-trivial exercise. This exercise is tricky even for seasoned ops and DevOps architects and engineers implementing traditional software solutions, because there is a fine line between a fantastic software lunch with lots of happy users and operators, and a disaster that results in performance and availability degradation, compliance and security nightmares, and integration disasters. While the following screenshot includes only sample data, it will nonetheless give readers a good idea of how we were analyzing real-life problems to determine the “State of Kubernetes in 2019.”
The EMA “The State of Kubernetes in 2019” research project aims to reveal exactly these examples to show what works well today and what Kubernetes users need to watch out for. But in addition to talking to hundreds of practitioners, this new research project is based on millions of data points derived from real-life support cases on mostly public forums, such as StackOverflow, Reddit, and Slack. We used a big data analytics approach to determine common problems of developers and IT operators and technology adoption patterns for both of these groups. While it would be elegant to tell readers that we leveraged machine learning and AI algorithms to reveal the interesting bits within our multiple TB-large data set, I will be honest and tell readers that while ML/AI helped us automate certain aspects of the analysis (e.g. selecting the best options for data visualization), there was much grunt work to do when it came to identifying variables and testing, revising, and retesting queries.
While the full report will be available for all of our lucky EMA subscribers, I can share numerous also very interesting charts that did not make the final report, mainly because they were too detailed.
You can find many more charts on the “State of Kubernetes in 2019” here and in my continuously updated practice newsstream on Wakelet.
