Feature #11664
open
provide rich output support for job execution
Description
As a user, when executing a job, I want to be able to be able to provide structured output, that can be used later on by Foreman.
For example, package upgrade - when upgrading existing packages, the output could be:
1. failed (already covered)
2. success -> which breaks down into two:
2a. nothing to upgrade - success
2b. upgraded nn packages, full ist of packages are [..]
This would allow us later on to have a more specific views for job result, for example:
1. overview page could provide hosts that were actually modified (or maybe even searched).
2. per job UI could be developed (surely for our locked templates, and maybe other 3rd customizations via community templates)
Updated by Ivan Necas about 9 years ago
It seems like it would make sense to extract the structured data into reports, side-by-side to puppet and other data?
Updated by Ivan Necas about 9 years ago
One of the most difficult things is probably how to extract the data form the runs, and maybe it makes sense to empover other tools to prepare the output (ansible/salt) and unify that in the Foreman at the end. In every case, it might be just processing the raw output (either the shell output in case of ssh, of some structured reprots from satl/ansible) and turning that into some common structure, ideally agreen on for several providers.
Updated by Ohad Levy about 9 years ago
Another thing this could lead into is chaining jobs, based on output, for example, workflow based on job output and potentially across hosts.