Five profiles, one evening, and a bug in our own summarizer
We ran five real Java Flight Recorder files through our own tool. All five said the application was idle. All five were wrong, and the reason was in our summarizer.
ReadWriting
What profiling real applications actually turns up — including when the thing it turns up is a bug in flamelens. Atom feed.
What the wild is fixing
Every profile analyzed on flamelens contributes one anonymous data point per finding: the kind of problem and the library it lived in. Nothing else — no code, no method names, no who. And because scans happen in tuning sessions (scan → fix → rescan), we can also count how often each finding kind actually gets resolved — evidence of what's fixable, not just what's common. Free for everyone, including the maintainers of these libraries.
| Problem kind | Library | Times seen | Resolved in-session |
|---|---|---|---|
| gc | jdk | 1 | 0 |
| io | thymeleaf-core | 1 | 0 |
| cpu | jdk | 1 | 0 |
| allocation | spring-core | 1 | 0 |
| cpu | node core | 1 | 0 |
Maintainers
If your library shows up here and you'd like the fuller anonymized breakdown, get in touch — this data is free to maintainers, always.