apiVersion: autoscaling/v2beta1
kind: HorizontalPodAutoscaler
metadata:
  name: myservice
  namespace: mynamespace
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: myservice
  minReplicas: 1
  maxReplicas: 3
  metrics:
  - type: Resource
    resource:
      name: memory
      targetAverageValue: 700Mi

After 15 min, i see REPLICAS grown up to 2. Since then it's not coming down.

C:\Users\myuser>kubectl get hpa
NAME            REFERENCE                  TARGETS           MINPODS   MAXPODS   REPLICAS   AGE
myservice    Deployment/myservice    490412032/700Mi          1         3         2          4h14m

There is no traffic to these services, as shown below MEMORY is well below 700Mi (as mentioned in HPA metrics).

C:\Users\myuser>kubectl top pod
NAME                                     CPU(cores)   MEMORY(bytes)
myservice-6ff6bdc8d-jx4pc             29m          463Mi
myservice-6ff6bdc8d-zktnm             29m          471Mi

I am puzzled why the 2nd replica has got created and why it's not going down ? Am i missingl something ? please suggest.