Synopsys:
1. Two network zones with firewall in between them, referenced as zone A and zone B.
2. Bitbucket Data Center v8.9.7 installed with 8 nodes in zone A.
3. Bitbucket mirrors installed and configured in zone A.
4. One Bitbucket mirror installed and configured in zone B, traffic is allowed only between Bitbucket DC in zone A and mirror in zone B. This enables users in zone B to fetch repositories created in zone A.
5. Jenkins deployed in zone A, working exclusively with zone A Bitbucket DC.
6. Jenkins properly configured with application link in Bitbucket and Bitbucket registered as a consumer in Jenkins.

Issues and observations:

1. Creating a multibranch pipeline using "atlassian-bitbucket-server-integration-plugin" won't allow us to select the server/mirror, even if only one mirror is unreachable for Jenkins (the one deployed in zone B).

2. Testing the connection will eventually time-out

3. Jenkins logs clearly show that only one mirror is unreachable. Other mirrors are reachable, but do not contain our repo, still not ok, as we have a very similar error message, e.g. "Failed to retrieve ...."

It looks like the issue is manifesting whenever a mirror from many is unreachable by Jenkins, so, I think the logic should be changed to present only available endpoints (main or mirrors) and not fail all if one is not replying.