Stale "Host connection failure" alarm remains in vCenter after an upgrade

Products

• VMware vCenter Server

Issue/Introduction

• After upgrading vCenter Server or ESXi hosts, a "Host connection failure" alarm is triggered.
• The ESXi host is fully functional, reachable via ping/SSH, and shows a "Connected" state in the vSphere Client.
• Attempting to "Reset to Green" on the triggered alarm does not clear the message.
• Restarting the vpxd service on vCenter Server does not resolve the stale alarm.

Environment

VMware vCenter Server 8.x

Resolution

 To clear the stale alarm state when standard resets fail, force a refresh of the alarm definition:

1. Log in to the vSphere Client.
2. Navigate to the vCenter Server.
3. Go to the Configure tab and select Alarm Definitions.
4. Locate the Host connection failure alarm 
5. Click Edit on the alarm definition.
6. Navigate through the edit wizard without making any changes to the triggers or actions.
7. Click Save.

Saving the definition forces vCenter to re-evaluate the status of all objects monitored by that alarm, which clears the stale "ghost" state.