Can you please elaborate steps to reproduce the issue?

I can't reproduce this.

Looks like it turns unhealthy if the remote harbor registry fails. Btw, I think the the unhealthy depends on ping rather than healthcheck.

Login to harbor portal, under Administration->Registries, add a a new endpoint of type harbor
In the form, select Provider “Harbor”, and fill all the required details. (Endpoint URL, etc..), Click “TEST CONNECTION” button to verify it is accessible and click “OK” to create it

After that endpoint has been created successfully, you would see in the new endpoint added to
Administration->Registries and its status is “Healthy”.

Up till now all good.

Now simulate an issue on the remote endpoint and make it inaccessible, (e.g: change replica of core to 0, or any other approach you want to make endpoint not reachable), as a result, the excepted behavior is that the “Status” column will be modify to “Unhealthy”, but it remain “Healthy” regardless if endpoint is accessible or not. (you can refresh the page to see it does make any changes)

Attaching a screenshot showing two endpoint one with Healthy, and second with Unhealthy (before my code change, both were Healthy regardless if there were accessible or not)

Even without the changes too. it stays the same. you can try reproducing this in the https://demo.goharbor.io