We are considering using chef to manage our application servers. So, we are creating recipes which configure resources for us as needed in certain situations.
But what happens if the execution fails? Do we have to manage for this cases manually? Are the logs of failing recipes posted somewhere for centralized monitoring? Are there common solutions for this cases like adding some kind of fail handler which notifies the interested parties?
UPDATE: We are running chef-client on the servers as a daemon with -d -i 60