RELEASED: Public: March 5, 2025
Resolved issues
- Unnecessary delays provisioning shared agents/clouds whose leases were recently returned (BEE-55432)
-
Previously, if a controller or a replica of an High Availability (HA) controller had leased all shared agents (or agents from a shared cloud) matching a given label, a different controller or replica of the same High Availability (HA) controller may be told by operations center that this label cannot be provisioned using the shared agent system.
This misleading information could be cached for up to five minutes even after some or all of the leases have been returned by the previous owner, blocking the new controller or replica from even asking operations center whether its outstanding queue items might now be satisfiable. The queue items would eventually be scheduled, but after an unnecessary delay during which the shared agents are idle, other controller or replica might unfairly lease them out of turn.
Now, the status of labels on the controller is tied only to the static configuration of the shared agents/clouds as defined on operations center. The current lease status is only considered when deciding whether to provide an agent.