There is a time in the life of every project when Database Administrator starts to look suspiciously at queries generated by our application. And then he (or she) sends us long list of queries that are not optimized, do not have proper indexes set or are executed multiple times in a very short time. And of course our role is to fix it.
Some issues are hard to solve but most of them are effect of our mistakes: N+1 problem in Hibernate is well-known (e.g. StackOverflow discussion) or checking the same condition (e.g. count something to figure visibility) multiple times for many components on the same page without caching it it any way. Of course we could turn on sql/hql logging during development process but sometimes we might not notice that here or there there is a problem with some query.
And here comes the solution. Simple parser analyzing hql logs and reporting which queries are executed too often. It’s far from being perfect or super universal but it works I called it Hibernate Smoke Detector.