Unable to specify SoftError for webhook providers

[krasoffski]

What happened:
Following PR #4166 introduced behavior when any not soft error leads to external-dns exit.

As a result, when webhook provider returns any error or non http.OK (code 200) external DNS restarts.
Here is example from code:
https://github.com/kubernetes-sigs/external-dns/blob/v0.14.1/provider/webhook/webhook.go#L181

Previous behavior was only log error and try next iteration.
This leads to problem with usage of buil-in webhook server and adding workaround for webhook provider which always returns 200 despite the errors to avoid external-dns retarts and possible crash loop.

What you expected to happen:
I expect that webhook provider errors is counted as soft errors or there is ability to distinguish (configure) fatal errors or soft errors will be returned by web-hook provider. Or configuration parameter which switchers web-hook errors as soft only errors.

How to reproduce it (as minimally and precisely as possible):
Just create webhook provider from examples which returns not 200 response in some cases. External-dns exits with error.

Environment:

• External-DNS version (use external-dns --version): v0.14.1
• DNS provider: from examples.

[SimonKienzler]

Does #4319 address your issue? It returns SoftErrors for 5xx errors on the webhook provider side. As far as I understand, 4xx errors are considered to be a problem on ExternalDNS's side, so no retry should be attempted. v0.14.2 contains the change.

[krasoffski]

@SimonKienzler
Thank you. I will check and come back with answer (close the issue or add more details).

[krasoffski]

Looks like what we need for web hook providers. Thank you!