[_visgean]

I have used that approach before with sentry. It was a non-issue. It depends on nature of the project of course, we had a system that was running every second so if it failed it generated a lot of data..