I spent over a decade as a consultant working for dozens of companies in many fields and pursuits. The diversity of each code base is tremendous. This article will try to define general rules for modernizing legacy code that would hopefully apply to all. But it comes from the angle of a Java developer. When writing this, my primary focus is on updating an old Java 6 era style J2EE code to the more modern Spring Boot/Jakarta EE code.

As systems get ever more complicated with more layers in the application stack, it can become more difficult to keep track of what is happening at all of the different, discrete levels and this is where distributed tracing comes into play.

When do you delete instrumentation? You delete instrumentation when you delete code. Other than that, if you’re doing things right: almost never. One of the best things about honeycomb is that it completely transforms the incentives around preserving instrumentation. With metrics-based tools, the most valuable metrics are always custom metrics. You need to define a custom metric for literally any question you might ever want to ask about the app and its utilization or performance.

Getting a good grasp on your application, especially when it is distributed across multiple clouds, kubernetes clusters and serverless functions is not an easy fit.

Any significant shift in an organization’s software engineering culture has the potential to feel tectonic, and observability (o11y for short)—or more specifically, Observability Driven Development—is no different. Leaning into observability, which calls for tool-enhanced investigation, hypothesis testing, and data richness can be cumbersome even for the most veteran of teams.