> On Aug 25, 2014, at 4:44 PM, Ramkumar R. Aiyengar <[email protected]> > wrote: > > I am in the process of looking at some of the ERROR level log output coming > from Solr to set up alarms, but currently the severity of what ERROR means is > kind of mixed across the code base. I am happy to fix this where I find, but > some guidance on what the various levels mean would be helpful. This is what > I would have expected: > > • ERROR: Shouldn't happen, indicates a bug or misconfiguration. Leads > to loss of functionality or some operation failing. Any occurrence indicates > something needs to be fixed. > • WARN: Recoverable problem, might genuinely happen in rare cases. If > it happens too often, might indicate an issue or misconfiguration. Bad input > data could fall into this category, or should it be INFO? > • INFO: Informational messages which would help in investigation, > indicates expected behaviour. > Let me know if this is not accurate.. >
Looks right overall. Which is not to say you won’t find an abuse here or there… bq. Bad input data could fall into this category, +1 I’ve been using more DEBUG as well. I think INFO should not spam (like our default successful add logging does) - it should just be useful always logged stuff to help with debugging and monitoring and operations. DEBUG can be a bit more spammy and just whatever is useful if developing in that area. - Mark http://about.me/markrmiller --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
