Currently Salt is using
supportconfig utility alongside with the SUSE Manager installation. Every single bugreport for Salt, that contains its logs are always useless that regard. It might work for SUSE Manager, but for Salt it doesn't. And even after asking user to switch the cluster into a debug mode or even trace mode, it is still far away from being useful enough, so the supporter don't need to dance with the tambourine around the fire, guessing what ghost just munged what component.
What if we can reuse Salt's internal tooling and develop some more tools that will not just grab logs/sysinfo from what is already there, but also get much more data, "x-ray" minions, turn them into a debug mode, collect right data, perform tests realtime etc?
For the record, I have no idea where we can go and how. Maybe someone wrote something similar that direction already?.. Come, join and R&D together!
Looking for mad skills in:
Nothing? Add some keywords!
This project is part of:
Hack Week 17
This project is one of its kind!