Apply new knative runners to remaining benchmarks

[dhschall]

In #238 we implement a new workflow that use Github runners for CI testing.
However, the workflow is not implemented for all benchmarks yet:

The benchmarks and the once net to be done:

• AES
• Authentication
• Apply new knative runners to: Chained-eventing #258
• Apply new knative runners to: Chained-serving #259
• Fibonacci
• Apply new knative runners to: corral #260
• Apply new knative runners to: gg-excamera #261
• Apply new knative runners to: gg-fibonacci #262
• Apply new knative runners to: gg-llvm #264
• Hotel-App #263
• Apply new knative runners to: map-reduce #265
• Online-shopping
• Apply new knative runners to: stacking-training #266
• trace-plotter
• Apply new knative runners to: tuning-halving #267
• Apply new knative runners to: video-analytics #268

What needs to be done
Before applying the service create the knative cluster using kind:

## Setup a Knative cluster to test the service
  - name: Create k8s Kind Cluster
    run: bash ./runner/scripts/01-kind.sh

  - name: Install Serving
    run: bash ./runner/scripts/02-serving.sh

  - name: Install Kourier
    run: bash ./runner/scripts/02-kourier.sh

  - name: Setup domain
    run: |
      INGRESS_HOST="127.0.0.1"
      KNATIVE_DOMAIN=$INGRESS_HOST.sslip.io
      kubectl patch configmap -n knative-serving config-domain -p "{\"data\": {\"$KNATIVE_DOMAIN\": \"\"}}"

Make sure to user this versions to be consistent

    env:
      KIND_VERSION: v0.14.0
      K8S_VERSION: v1.23

Note: As the scripts only install kubectl but not kn one can only use kubectl commands to apply and interact with services.
Either add kn installation or use the respective kubectl command instead of kn. I.e. kubectl apply -f <> instead of kn service apply -f <>

[ustiugov]

@dhschall please create children Issues and assign them to the corresponding maintainers

[ustiugov]

apparently, there is a workaround to run Eventing (and probably all other more complex knative setups) that removes all resource requests to k8s (which may not be possible to satisfy when running on a single GH runner with 3 CPU cores and little memory)

kubectl get deployment -n knative-serving -o yaml | yq 'del(.items[]?.spec.template.spec.containers[]?.resources)' | kubectl apply -f -
kubectl get deployment -n knative-eventing -o yaml | yq 'del(.items[]?.spec.template.spec.containers[]?.resources)' | kubectl apply -f -

As suggested in csantanapr/knative-kind#47 (comment).