The second feature is more a generic retry/restart approach. I would like to be able to retry a failed/aborted job using the user interface. For instance, if I have a build job that was aborted during the compilation, I would like to click on a restart button (see snapshot below) and select the job/step that I want to retry. If the job succeeds after the retry, then its dependencies are triggered exactly as if it never failed. We often have network glitches (out of space, connection lost) or incremental build errors that lead to a build failure. When you have 20 different targets to build and only one fails because the machine ran out of space, we often retry the job after a manual cleanup. It is a very useful feature.
 
The following snapshot is the restart feature in the product AnthillPro that we are currently using.