Currently, when IPA spawns a tool to perform a long-running task on the node (like wiping out the storage on node clean up), there is no way for the operator to figure out the progress of that long-running operation (is it still running or maybe hung?) as well as to have an estimation on when it might finish. Possible solution can be to capture the information at the standard output streams of a spawned tool, buffer that and eventually deliver the data to the Ironic conductor (probably piggy-backing the heartbeats) for further consumption by the cli or web-ui clients.
Proposed spec: https://review.openstack.org/492702
Moving out of OSP-15, bumping to 16.
I'm marking closed/deferred as the community pushed back on this effort and deploy_steps in Victoria at least answers the "what exactly is going on" portion of what this sought to answer.