[core] Incorrect resource detection by accounting thread inside MCAD

[asm582]

0805 12:21:55.464941   80515 cache.go:223] Total capacity cpu 0.00, memory 0.00, GPU 0, used cpu 0.00, memory 0.00, GPU 0, free space cpu 0.00, memory 0.00, GPU 0

The accounting thread runs every 1 sec and it is observed that sometimes at bootstrap time the accounting is incorrect. This will cause the target AW to be put in the back-off queue for another 20 secs, which will cause a delay of dispatch about ~22 seconds for the same AW.

FYI the cluster has below nodes:

(base) abhishekmalvankar@Abhisheks-MBP multi-cluster-app-dispatcher % oc get nodes        
NAME                                         STATUS   ROLES                  AGE   VERSION
ip-10-0-148-245.us-east-2.compute.internal   Ready    worker                 51d   v1.26.3+b404935
ip-10-0-153-245.us-east-2.compute.internal   Ready    worker                 30d   v1.26.3+b404935
ip-10-0-156-1.us-east-2.compute.internal     Ready    control-plane,master   51d   v1.26.3+b404935
ip-10-0-159-143.us-east-2.compute.internal   Ready    worker                 30d   v1.26.3+b404935

[asm582]

cache has been simplified and the above error does not occur when tested manually using a debugger on the local machine.