We build services, potentially used by hundreds, thousands or even millions of people. And despite all the testing we do, some operations and interactions with these services will not work out the way we expect or hope. Wouldn’t it be great to be able to reconstruct what led to a problem? And then analyze if the same problem occurred at other times?
In this talk, I want to look at structured logging. What are the benefits of it over plain-text logging? How can it be used effectively?
Logging, by itself, can help to trace bugs. Structured logging adds additional, post-logging capabilities when used correctly. Because log events are structured data, they can be analyzed much more efficiently and effectively than plain text log messages. Furthermore, one can even build graphs and diagrams to visualize what’s happening in a service nicely.