k8s-crd-benchmark
overview
This is meant as a informal benchmark for the k8s API server in regards to creating and modifying CRDs.
usage
installation
k3d cluster create -p "30000:[email protected]:0" -p "32000:[email protected]:0" --image rancher/k3s:v1.21.8-k3s2 -v /dev/mapper/nvme0n1p3_crypt:/dev/mapper/nvme0n1p3_crypt flyte
kubectl create -f k8s/service-account.yaml
kubectl create -f k8s/prometheus.yaml
kubectl create -f k8s/grafana.yaml
resources
deploying prometheus / grafana on k8s
https://sysdig.com/blog/monitor-kubernetes-api-server/ https://medium.com/@gurpreets0610/deploy-prometheus-grafana-on-kubernetes-cluster-e8395cc16f91 https://blog.freshtracks.io/a-deep-dive-into-kubernetes-metrics-part-4-the-kubernetes-api-server-72f1e1210770 metrics overview grafana board
update vs patch
https://blog.atomist.com/kubernetes-apply-replace-patch/
benchmarks
50
patch single 8.257, 8.261, 8.235 patch all 0.067, 0.102, 0.039 update single 8.266, 8.238, 8.239 update all 0.055, 0.042, 0.038
