Hotfix - v0.1.3

Make sure there's always one entry per dns, so the DNS is not entirely removed.

v0.1.2

This version contains minor improvements, to make the service more stateless.

Fixed:
TcpProbe timeout is no longer using the host-system default. It uses the configured timeout for the specific check.

Feature:
During startup routing the service will check for existing DNS records and compare it to its own targets.
If the service was restarted during a downtime it would not have recovered the previous unavailable state, but now it recovers the state from DNS and will readd the entry as soon as it has a succeeded check on the unavailable target.

Initial Release

I had to fix the workflow for the tag-builds first.