How the World’s Leading Companies Choose How and What to Log (Based on analysis of GitHub’s top 400,000 repositories)
It’s no secret that log files are noisy and difficult to navigate. With seemingly endless data coming in from our applications, the uses for such data must also be endless. Or so we thought. The abundance of log aggregation and analyzer tools, gives us the ability to parse the data for security, BI and application monitoring purposes.
BUT, the value of our logs is entirely dependent on what, and how, we write statements to the logs in the first place.
So, we set out to understand how developers use logs by looking at 779,236 logging statements on GitHub. What went in? What was left out? And what does all this mean for the value of our log files?
- ERROR, WARN or FATAL? – An examination of the log level breakdown use for the average Java application
- Is Standard Java Logging Dead? – Putting the focus on the use of standard java.util.logging levels versus more popular frameworks like Log4j, Log4j2, and Logback
- Over 50% of Java Logging Statements Are Written Incorrectly – A look at the crucial missing piece that prevents the logs from pointing to the real root cause of your errors
- What’s the Top Java Logging Method on GitHub? – Parameters, concatenations or both; which logging method are your peers using, and which method is right for you?