Why does the istio-proxy sidecar have so high default limits?

Contributors Operator
1

We have installed istio-operator into our kubernetes cluster.

By default, the istio sidecars (istio-proxy) are assigned the following request and limits:

    resources:
      requests:
        cpu: 100m
        memory: 128Mi
      limits:
        cpu: 2000m
        memory: 1024Mi

Normally, the limits of containers is reasonable and typically only double the requests, unless it is a service that spikes suddenly.

I am investingating in prometheus the limits and requests of all our environment. We have tons of containers and the istio-proxy, with their huge limits compared to usage, makes the cluster have huge limit.

Isent having that high limits messing with kube-scheduler making it add less pods to each node due to the limit being so high?

Is there any reason to not change it to something like this?

    resources:
      requests:
        cpu: 100m
        memory: 128Mi
      limits:
        cpu: 200m
        memory: 256Mi