We often work on large websites with strict change control practices and scheduled release cycles. Sometimes we also hand over the systems to the client for production and don’t have direct access ourselves.

Some bugs have a nasty habit of only occurring in production, this may be due to high load, odd/old browsers, changes in data, or just because test scenarios don’t cover every eventuality.

What this means is that when we have a bug in production we can only understand it through the error logging we have already built into the system. If we need to put in place additional logging we usually lose the chance of actually fixing the bug for another release cycle.