Run Conformance and Performance tests with github actions

.github/actions/install-dependencies/action.yaml

@@ -0,0 +1,20 @@
+name: InstallDependencies
+description: 'Installs Go, Docker, Ginkgo, EKSCTL binaries'
+runs:
+  using: "composite"
+  steps:
+    - uses: actions/setup-go@v4
+      with:
+        go-version: "1.20"
+        # go-version-file: go.mod
+        check-latest: true
+    - name: Set up ginkgo
+      shell: bash
+      run: |
+        # Install ginkgo version from go.mod
+        go install -mod=mod github.com/onsi/ginkgo/v2/ginkgo
+    - name: Set up eksctl
+      shell: bash
+      run: |
+        curl --silent --location "https://github.com/weaveworks/eksctl/releases/latest/download/eksctl_$(uname -s)_amd64.tar.gz" | tar xz -C /tmp
+        sudo mv /tmp/eksctl /usr/local/bin/

.github/workflows/e2e-conformance.yaml

@@ -0,0 +1,39 @@
+name: e2e-conformance-tests
+
+on:
+  workflow_dispatch: {}
+  schedule:
+    - cron: "0 0 * * *" # Run Everyday at Midnight
+
+permissions:
+  id-token: write
+  contents: read
+
+jobs:
+  e2e-conformance-tests:
+    strategy:
+      fail-fast: false
+      matrix:
+        ip-family: [ IPv4, IPv6 ]
+        # kubernetes-versions: ["1.25", "1.26", "1.27"]
+    if: github.repository == 'aws/aws-network-policy-agent'
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout latest commit in the PR
+        uses: actions/checkout@v3
+      - name: Install Dependencies
+        uses: ./.github/actions/install-dependencies
+      - uses: aws-actions/configure-aws-credentials@v1
+        with:
+          role-to-assume: ${{ secrets.OSS_ROLE_ARN }}
+          aws-region: ${{ secrets.AWS_DEFAULT_REGION }}
+          role-duration-seconds: 14400
+      - name: Run e2e conformance test
+        env:
+          RUN_CONFORMANCE_TESTS: true
+          KUBERNETES_VERSION: 1.27
+          CNI_ADDON_VERSION: v1.13.3-eksbuild.1
+          CNI_ADDON_CONFIGURATION: ""
+          IP_FAMILY: ${{ matrix.ip-family }}
+        run: |
+          ./scripts/run-tests.sh

.github/workflows/performance-tests.yaml

@@ -0,0 +1,41 @@
+name: Performance tests
+
+on:
+  workflow_dispatch: {}
+  schedule:
+    - cron: "0 9 * * 2" # every Tuesday
+
+permissions:
+  id-token: write
+  contents: read
+
+jobs:
+  performance-tests:
+    strategy:
+      fail-fast: false
+      matrix:
+        ip-family: [ "IPv4", "IPv6"]
+        # kubernetes-versions: ["1.25", "1.26", "1.27"]
+    if: github.repository == 'aws/aws-network-policy-agent'
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout latest commit in the PR
+        uses: actions/checkout@v3
+      - name: Install Dependencies
+        uses: ./.github/actions/install-dependencies
+      - uses: aws-actions/configure-aws-credentials@v1
+        with:
+          role-to-assume: ${{ secrets.OSS_ROLE_ARN }}
+          aws-region: ${{ secrets.AWS_DEFAULT_REGION }}
+          role-duration-seconds: 14400
+      - name: Run performance tests
+        env:
+          RUN_PERFORMANCE_TESTS: true
+          KUBERNETES_VERSION: 1.27
+          NODES_CAPACITY: 10
+          INSTANCE_TYPE: c5.xlarge
+          CNI_ADDON_VERSION: v1.13.3-eksbuild.1
+          CNI_ADDON_CONFIGURATION: ""
+          IP_FAMILY: ${{ matrix.ip-family }}
+        run: |
+          ./scripts/run-tests.sh

scripts/README.md

@@ -0,0 +1,12 @@
+## Integration Test scripts
+
+This package contains shell scripts and libraries used for running e2e integration tests.
+
+### run-test.sh
+
+`run-test.sh` can run various integration test suites against the current revision in the invoking directory. 
+
+#### Tests
+The following tests are valid to run, and setting the respective environment variable to true will run them:
+1. Conformance Tests - `RUN_CONFORMANCE_TESTS`
+2. Performance Tests - `RUN_PERFORMANCE_TESTS`

scripts/lib/cleanup.sh

@@ -0,0 +1,24 @@
+
+function check_path_cleanup(){
+
+    local worker_nodes=$(kubectl get nodes -o custom-columns=NAME:.metadata.name --no-headers)
+    for node in $worker_nodes
+    do
+        export NODE=$node
+        envsubst '$NODE' < ${DIR}/test/check-cleanup-pod.yaml > ${DIR}/test/check-cleanup-pod-$node.yaml
+        kubectl apply -f ${DIR}/test/check-cleanup-pod-$node.yaml
+        rm -rf ${DIR}/test/check-cleanup-pod-$node.yaml
+    done
+    sleep 20
+
+    for node in $worker_nodes
+    do
+        if [[ $(kubectl get pods $node -ojsonpath="{.status.phase}") == "Failed" ]]; then
+            echo "BPF files not cleaned up on $node.. $(kubectl logs $node)"
+            exit 1
+        fi
+        kubectl delete pods $node
+    done
+
+    echo "BPF files were cleaned up from the nodes"
+}

scripts/lib/cloudwatch.sh

@@ -0,0 +1,40 @@
+function install_cloudwatch_agent(){
+
+    local perf_cluster_name=""
+    if [[ $IP_FAMILY == "IPv4" ]]; then
+        perf_cluster_name="eks-network-policy-perf-v4"
+    else
+        perf_cluster_name="eks-network-policy-perf-v6"
+    fi
+
+    echo "Create IAM Service Account for CW agent"
+    kubectl create ns $CW_NAMESPACE
+
+    eksctl create iamserviceaccount \
+        --cluster $CLUSTER_NAME \
+        --name cloudwatch-agent \
+        --namespace $CW_NAMESPACE \
+        --attach-policy-arn $CW_POLICY_ARN \
+        --approve
+
+    echo "Install Cloudwatch Agent DS"
+    kubectl apply -f https://raw.githubusercontent.com/aws-samples/amazon-cloudwatch-container-insights/latest/k8s-deployment-manifest-templates/deployment-mode/daemonset/container-insights-monitoring/cwagent/cwagent-serviceaccount.yaml
+
+    echo '{ "logs": { "metrics_collected": { "kubernetes": { "metrics_collection_interval": 30, "cluster_name": "'${perf_cluster_name}'" }},"force_flush_interval": 5 }}' | jq > cwagentconfig.json
+    kubectl create cm -n $CW_NAMESPACE cwagentconfig --from-file cwagentconfig.json
+    kubectl apply -f https://raw.githubusercontent.com/aws-samples/amazon-cloudwatch-container-insights/latest/k8s-deployment-manifest-templates/deployment-mode/daemonset/container-insights-monitoring/cwagent/cwagent-daemonset.yaml
+
+    # Allow CW agent to startup and push initial logs
+    sleep 60
+}
+
+function uninstall_cloudwatch_agent(){
+
+    eksctl delete iamserviceaccount \
+        --cluster $CLUSTER_NAME \
+        --name cloudwatch-agent \
+        --namespace $CW_NAMESPACE || echo " IAM Service Account role not found"
+
+    rm -rf cwagentconfig.json || echo "CW agent config not found"
+    kubectl delete namespace $CW_NAMESPACE || echo "No namespace: $CW_NAMESPACE found"
+}

scripts/lib/cluster.sh

@@ -0,0 +1,63 @@
+
+
+function set_cluster_defaults(){
+
+    CLUSTER_NAME=network-policy-${RANDOM}
+    : "${AWS_REGION:=us-west-2}"
+    : "${AMI_FAMILY:=AmazonLinux2}"
+    : "${NODEGROUP_TYPE:=linux}"
+    : "${NODES_CAPACITY:=3}"
+    : "${INSTANCE_TYPE:=t3.large}"
+    : "${KUBERNETES_VERSION:=1.27}"
+    : "${IP_FAMILY:=IPv4}"
+    : "${CNI_ADDON_VERSION:=v1.13.3-eksbuild.1}"
+    : "${CNI_ADDON_CONFIGURATION:=""}"
+    : "${CW_NAMESPACE:=amazon-cloudwatch}"
+    : "${CW_POLICY_ARN:=arn:aws:iam::aws:policy/CloudWatchAgentServerPolicy}"
+    : "${NETWORK_POLICY_NS:=netpol-test}"
+    : "${ENDPOINT_URL:=""}"
+}
+
+function create_cluster(){
+
+    cat <<EOF > eks-cluster.yaml
+    apiVersion: eksctl.io/v1alpha5
+    iam:
+        withOIDC: true
+    addons:
+      - name: vpc-cni
+        version: ${CNI_ADDON_VERSION}
+        configurationValues: ${CNI_ADDON_CONFIGURATION}
+      - name: coredns
+      - name: kube-proxy
+    kind: ClusterConfig
+    kubernetesNetworkConfig:
+        ipFamily: ${IP_FAMILY}
+    managedNodeGroups:
+      - amiFamily: ${AMI_FAMILY}
+        desiredCapacity: ${NODES_CAPACITY}
+        instanceType: ${INSTANCE_TYPE}
+        labels:
+            alpha.eksctl.io/cluster-name: ${CLUSTER_NAME}
+            alpha.eksctl.io/nodegroup-name: ${CLUSTER_NAME}-${NODEGROUP_TYPE}-nodes
+        maxSize: ${NODES_CAPACITY}
+        minSize: 1
+        name: ${CLUSTER_NAME}-${NODEGROUP_TYPE}
+        tags:
+            alpha.eksctl.io/nodegroup-name: ${CLUSTER_NAME}-${NODEGROUP_TYPE}-nodes
+            alpha.eksctl.io/nodegroup-type: managed
+    metadata:
+        name: ${CLUSTER_NAME}
+        region: ${AWS_REGION}
+        version: "${KUBERNETES_VERSION}"
+EOF
+
+    eksctl create cluster -f ./eks-cluster.yaml
+}
+
+function delete_cluster(){
+
+    eksctl delete cluster -f ./eks-cluster.yaml
+    rm -rf ./eks-cluster.yaml
+}
+

scripts/lib/network-policy.sh

@@ -0,0 +1,72 @@
+
+function install_network_policy_mao(){
+
+    local options=" --no-cli-pager"
+    if [[ ! -z $ENDPOINT_URL ]]; then
+        options+=" --endpoint-url $ENDPOINT_URL"
+    fi
+
+    if [[ ! -z $CNI_ADDON_CONFIGURATION ]]; then
+        options+=" --configuration $CNI_ADDON_CONFIGURATION"
+    fi
+
+    aws eks create-addon \
+        --addon-name vpc-cni \
+        --addon-version $CNI_ADDON_VERSION \
+        --resolve-conflicts overwrite \
+        --cluster-name ${CLUSTER_NAME} $options
+
+    local status=""
+    local retries=30
+    local try=0
+    while [[ $status != "ACTIVE" && $try -lt $retries ]]
+    do
+        status=$(aws eks describe-addon \
+            --addon-name vpc-cni \
+            --cluster-name ${CLUSTER_NAME} $options | jq -r '.addon.status')
+        echo "Addon status - $status" 
+        try=$((try+1))
+        sleep 10
+    done
+
+    if [[ $status != "ACTIVE" ]]; then
+        local err_message=$(aws eks describe-addon \
+            --addon-name vpc-cni \
+            --cluster-name ${CLUSTER_NAME} $options | jq -r '.addon.health')
+        echo $err_message
+        exit 1
+    fi
+
+    echo "Addon installed Successfully"
+}
+
+function install_network_policy_helm(){
+
+    echo "Installing Network Policy using VPC-CNI helm chart"
+    helm repo add eks https://aws.github.io/eks-charts
+
+    if [[ $IP_FAMILY == "IPv4" ]]; then
+        ENABLE_IPv4=true
+        ENABLE_IPv6=false
+        ENABLE_PREFIX_DELEGATION=false
+    else
+        ENABLE_IPv4=false
+        ENABLE_IPv6=true
+        ENABLE_PREFIX_DELEGATION=true
+    fi
+
+    helm upgrade --install aws-vpc-cni eks/aws-vpc-cni --wait --timeout 600 \
+        --namespace kube-system \
+        --set enableNetworkPolicy=true \
+        --set init.env.ENABLE_IPv6=$ENABLE_IPv6 \
+        --set image.env.ENABLE_IPv6=$ENABLE_IPv6 \
+        --set nodeAgent.enableIpv6=$ENABLE_IPv6 \
+        --set image.env.ENABLE_PREFIX_DELEGATION=$ENABLE_PREFIX_DELEGATION \
+        --set image.env.ENABLE_IPv4=$ENABLE_IPv4
+
+}
+
+function install_network_policy_manifest(){
+
+
+}

scripts/lib/tests.sh

@@ -0,0 +1,16 @@
+function run_cyclonus_tests(){
+
+    kubectl create ns $NETWORK_POLICY_NS
+    kubectl create clusterrolebinding cyclonus --clusterrole=cluster-admin --serviceaccount=$NETWORK_POLICY_NS:cyclonus
+    kubectl create sa cyclonus -n $NETWORK_POLICY_NS
+
+    kubectl apply -f ${DIR}/test/cyclonus-config.yaml -n $NETWORK_POLICY_NS
+
+    kubectl wait --for=condition=complete --timeout=240m -n $NETWORK_POLICY_NS job.batch/cyclonus || echo "Job timed out after 4 hrs"
+    kubectl logs -n $NETWORK_POLICY_NS job/cyclonus > ${DIR}/results.log
+
+}
+
+function run_performance_tests(){
+    run_cyclonus_tests
+}

scripts/run-tests.sh

@@ -0,0 +1,39 @@
+#! /bin/bash
+
+set -Eeuox pipefail
+
+DIR=$(cd "$(dirname "$0")"; pwd)
+
+source ${DIR}/lib/cleanup.sh
+source ${DIR}/lib/cloudwatch.sh
+source ${DIR}/lib/cluster.sh
+source ${DIR}/lib/network-policy.sh
+source ${DIR}/lib/tests.sh
+
+: "${RUN_PERFORMANCE_TESTS:=false}"
+: "${RUN_CONFORMANCE_TESTS:=false}"
+
+cleanup() {
+
+  if [[ $RUN_PERFORMANCE_TESTS == "true" ]]; then
+    uninstall_cloudwatch_agent
+  fi
+
+  delete_cluster
+}
+
+trap cleanup EXIT
+
+set_cluster_defaults
+create_cluster
+
+if [[ $RUN_PERFORMANCE_TESTS == "true" ]]; then
+  echo "Runnning Performance tests"
+  install_cloudwatch_agent
+  run_performance_tests
+elif [[ $RUN_CONFORMANCE_TESTS == "true" ]]; then
+  echo "Running Conformance tests"
+  run_cyclonus_tests
+fi
+
+check_cleanup