I have a test cluster that has just a single workload running that is basically doing nothing. Our Grafana metric show that the 3-node cluster has been running at around 7% average utilization across the nodes.
When we enabled Istio in 2.3.2, we saw the utilization jump to around 11%. This is with only enabling Istio from the Rancher UI. We have not done any sidecar injection yet. Is this normal to see a 4% CPU jump when Istio is not doing any real work?
When I look at the pod CPU usage, before we enabled Istio, the pods were using between .001 and .013 percent usage with most of them at or below .003. After enabling Istio, they are using between .002 and .021 percent, with most of them at .005 or above.