memory and cpu metrics on the HPA lower than real memory and cpu usage

[medialearnuxx]

hi, i apologize if maybe this question or kind of issue was raise by others or not. but i have 6 micro services application that run on top of kubernetes. i implement linkerd to load balance the traffic to the pod. but when i compare the services/deployment.
at first, i got reported from team if HPA on the services that injected linkerd not worked. then i add these annotations to the deployment, config.linkerd.io/proxy-cpu-request and config.linkerd.io/proxy-memory-request. the HPA issues solved. but i found another issues that the memory and cpu metrics are mismatch because i also monitor my application services using grafana.
thanks

[medialearnuxx]

my pod resources
resources:
limits:
cpu: 500m
memory: 800M
requests:
cpu: 500m
memory: 800M

annotations:
config.linkerd.io/proxy-cpu-request: 500m
config.linkerd.io/proxy-memory-request: 800M