The classic statement, “it works great in my environment”, represents a lack of observability. Traditionally, operations and development teams have used different tools to observe the behavior of their systems. Operations have historically been largely limited to black-box monitoring. Ops could monitor the operating system, processes, memory, CPU, disk I/O, network I/O, some log files, and the database, but they had little visibility into the inner workings of the applications. When a system underperformed, the Ops team could confirm the fact that is was underperforming, but could not necessarily state why it was underperforming.
Eventually, the development team was brought in to investigate the problem in a lower environment, where they can use profilers and debuggers to perform white-box testing and potentially diagnose the problem and propose a solution. However, lower environments are out of context and often cannot reliably reproduce the problem. Regardless, the traditional...