add v1 model and tests for poddisruptionbudgets #108
Conversation
j18e
force-pushed
the
add-pod-disruption-budget
branch
from
bfe1e14 to
87054c6
Compare
j18e
force-pushed
the
add-pod-disruption-budget
branch
from
87054c6 to
c55591b
Compare
| values = ListField(six.text_type) | ||
|
|
||
|
|
||
| class LabelSelector(Model): |
There was a problem hiding this comment.
I see that there are three other LabelSelector model classes which seem to represent the same API resource already:
- https://github.com/fiaas/k8s/blob/master/k8s/models/deployment.py#L28
- https://github.com/fiaas/k8s/blob/master/k8s/models/job.py#L26
- https://github.com/fiaas/k8s/blob/master/k8s/models/pod_disruption_budget.py#L33
Perhaps one step towards improving this situation could be to move this LabelSelector class into a new meta_v1 module (to follow the naming convention) and use that here. At some point it would be good to switch to the same class in every model that uses LabelSelector, but since that is a breaking change I don't think it is necessary to do in this changeset. What do you think?
There was a problem hiding this comment.
i think it's a good idea, though we already have a common package which currently houses metav1 type objects. So i think it would go best in there. If we really want to call it meta_v1 we should move everything over there
|
The code analysis issues can be fixed by disabling the check for the test class; see for example https://github.com/fiaas/k8s/blob/master/tests/k8s/test_apiextensions_v1_custom_resource_definition.py#L31 |
This PR adds poddisruptionbudgets from the policy/v1 apiversion. PDB's were partially implemented with the v1beta1 version in 2018, but these were never put into use.