This isn't related to the sizes of your databases, but rather the wider growth of systems, applications, platforms, and even teams in your organization. In even the smallest companies, there can be an ever expanding list of things to support and manage. Developers aren't immune either, with ever growing numbers of packages, libraries, interfaces, APIs, and more that they might struggle to master.

The cloud adds another dimension as well. Vendors continue to update their systems, which means that your knowledge, scripts, and processes might become out of date without you changing anything. On top of that, the ways in which you manage and tune systems can shift underneath you, necessitating the need to re-learn skills on a platform that you might never actively upgrade or change.

The growing complexity and challenges can lead many technologists to minimize change, and try to limit the ways in which they look at systems. A recent article finds that many teams are being overloaded and not adapting well to the environments. Lots of staff uses tools they are familiar with, in familiar ways, and may not really be observing much of what is happening.

I can see this even in a singular SQL Server instance. While Microsoft has built a number of tools to help us better understand what is happening on the platform, there is great complexity, an ever growing number of DMVs, and even different interfaces that DBAs and developers may not take advantage of A simple example is Extended Events. While Trace/Profiler has been deprecated and doesn't capture a lot of what happens in modern platforms, many people continue to use those tools over XEvents for troubleshooting. Even as I experimented with both, and found the XEvent Profiler to be quicker to get using than Profiler, it more cumbersome and I struggle to think of using it first.

Old habits die hard, and while many of us cling to them, we're not really taking advantage of the modern IT infrastructure. There are new tools and new ways of using them that can help us better understand our environments. While there might be better AI assistance in the future, I don't hold my breath those will solve all the problems. At the very least, however, we ought to be slowly experimenting with better ways of managing our systems and taking advantage of new tools that might give us greater leverage to observe and manage increasingly larger numbers of systems.