v0.7.0 New arguments, new modes, metrics

This release is a massive refactoring of the code and deprecation of the old hour-based CLI arguments in favor of duration-based.
It's now possible to specify for how long exactly you want to keep your jobs or pods.
Instead of a minimal unit of 1h - -keep-successful=1 , you now can specify -delete-successful-after=5m.
It also fixes old bugs like duplication of events, when the same job or pod was showing in the logs twice.

boolen flag -legacy-mode is responsible for running the operator in either old mode or new one. Default mode is set to true, which means that operator will work as before, using -keep-successful flags.

To try new flags run the operator with -legacy-mode=false.

Changelog

c40354b Collect and expose prometheus metrics #42
daec120 fix legacy podController not being started
5ae2dee Add support for deletion of evicted pods, refactor
87df0bf New arguments format, new controller, deprecation warning
aa02a27 Upgrade go to 1.14, update dependencies

Docker images

• docker pull quay.io/lwolf/kube-cleanup-operator:v0.7.0
• docker pull quay.io/lwolf/kube-cleanup-operator:latest