Kuberhealthy gets stuck when a khcheck object is misconfigured

[jonnydawg]

Describe the bug
The kuberhealthy master will run into an InternalError when looking for khc changes:

Error when watching for khcheck changes: InternalError

Related to #222

Steps To Reproduce
Misconfigure any khcheck installed. An easy way to do this is to remove the - under podSpec.containers.

I think this issue is caused by the mismatching types here. Kuberhealthy expects podSpec.containers to be an array, but removing this - causes podSpec.containers to become and object instead of an array of objects.

Expected behavior
Misconfigure a check's podSpec.containers and observe that checks will no longer be scheduled.

Screenshots

time="2020-07-13T18:11:10Z" level=error msg="Error listing check configuration resources khcheckcrd.KuberhealthyCheckList.Items: []khcheckcrd.KuberhealthyCheck: khcheckcrd.KuberhealthyCheck.Spec: khcheckcrd.CheckConfig.PodSpec: v1.PodSpec.Containers: []v1.Container: decode slice: expect [ or n, but found {, error found in #10 byte of ...|tainers\":{\"image\":\"d|..., bigger context ...|0123456789ab\"},\"spec\":{\"podSpec\":{\"containers\":{\"image\

Versions

• Cluster OS: Debian
• Kubernetes Version: 1.11.7
• Kuberhealthy v2.2.0

[jonnydawg]

Currently dealing with this issue from the webhook server:

2020/07/18 03:18:19 http: TLS handshake error from 192.168.65.3:54690: remote error: tls: bad certificate
2020/07/18 03:18:49 http: TLS handshake error from 192.168.65.3:54944: remote error: tls: bad certificate