pyinfra-dev
List failed hosts on run
Is your feature request related to a problem? Please describe
When I have an operation fail on x amount of hosts, it doesn't allow me to run that same operation on only those failed hosts plus the ones it did not get to if it hit the failure limit easily.
Describe the solution you'd like
If I have 10 hosts, and my operations succeed on the first 3 and fail on the next 3, and my failure limit is 30%, I would like to have an option or some way to get a new inventory file to run those operations again on those failed hosts plus the inventory it did not have a chance to reach. Currently I believe the only way to do this is rerun the operations on all hosts again redundantly.
I like this idea a lot, similar to Ansible retry files. Could go one step further and skip any successful operations on the failed hosts as well.
