Just to be clear, the backoff retry handler is just the default one, and providers (several do) may override the handlers with custom ones to provide more fine-grained control when retrying upon service specific failures. They don't even have to extend the backoff limited retry handler. If we believe (I do) that this is a sane default behavior, I think we'd rather move the check up a level and don't force every retry handler to take care of that (and if it makes sense it is something we can easily make configurable per provider, if at any point we find it makes sense).
--- Reply to this email directly or view it on GitHub: https://github.com/jclouds/jclouds/pull/744#issuecomment-101316686
