vCenter 8.x
STS shares 150 worker threads with other modules. Each login attempt against a dead DC held a worker thread before timing out. With enough concurrent attempts, the 150-thread pool filled. Once the pool was exhausted, local system logins and internal authentication requests were forced to queue behind the pending AD requests, causing them to time out.
Applmgmt needs a quick SSO round-trip at startup to verify its own privileges (depending on sts). With SSO starved, that call didn't return, its health probe failed, vMon restarted it and the restart hit the same stuck SSO.
1. Check AD DNS record and /etc/hosts file
2. Validate Network Connectivity: curl -v telnet://<DC_FQDN>:389 (or 636 for LDAPS) from the VCSA shell to confirm the path is open.