- research for horizontal scaling
- Evan Drake
- Benjamin Serrato
- Steven Mendez
- Ryan Hoofard
- HPA: kubecatl get hpa [autoscaler name *optional] --watch[foreground watcher]
- Deployment: kubectl deploy [deployment name *optional] --watch [foreground watcher]
- findout what pods services daemonsets deployments and replicasets are doing
kubectl -n kube-system get all
- create a kind cluster with config as config.yaml
kind create cluster --config=k8s-files/config.yaml --name scaling
- create rust container
docker build . -t rust-demo
- load docker container into the kind cluster
kind load docker-image rust-demo --name scaling
- apply yaml files to the cluster and watch benny's hard work
kubectl apply -f scale-namespace.yaml
kubectl apply -f autoscale.yaml
kubectl apply -f components.yaml
kubectl apply -f deployment-rust.yaml
kubectl apply -f deployment-rust.yaml
- change default namespace with benny's amazing script
bash .namespace.sh namespace scale-namespace
- make sure you can curl and get an output
curl localhost:8080
- run create script in each metrics sub directory
- alias for this in .bashrc
- portforward grafana
kubectl port-forward grafana-6648db4b8b-cr5xx 3000:3000 -n monitoring
- alias for this in .bashrc
- grapha needs to be initialzed with prometheus at start
- import dashboard 1860
- select prometheus from the drop down near the bottom
- test with hey tool
hey -n 1000000 -z 5s http://127.0.0.1:8080