Agent fails to update task-level resource control

[amaltaro]

Impact of the bug
WMAgent

Describe the bug
While inserting T3_UK_ScotGrid_GLA into the RelVal agent, I noticed that only the site level resources were getting reflected from changes made at the SSB grafana view. So these slots were getting updated accordingly:

T3_UK_ScotGrid_GLA - 188 running, 1415 pending, 2000 running slots total, 1500 pending slots total, Site is Normal:

but not the task-level ones, example (which is now correct):

  Cleanup - 0 running, 0 pending, 25 max running, 25 max pending, priority 1
  Production - 0 running, 1400 pending, 2000 max running, 1400 max pending, priority 0

How to reproduce it
Change the production slots for a given site in Grafana SSB.

Expected behavior
In addition to the site-level pending and running slots, agents need to update the task-level thresholds accordingly as well.

This information is stored in the local relational database (which we usually refer to as resource-control) and it's automatically updated by the AgentStatusWatcher component (ResourceControlUpdater thread)

Additional context and error message
None