Merge branch 'main' into tudorman/gpu-tensor-integ-tests

Author: tudor-manea

.gitignore

@@ -11,4 +11,7 @@ terraform.*
 /test/**/final_*.yml
 coverage.txt
 generator/resources/*complete*.json
-terraform/eks/e2e/helm-charts
+terraform/eks/e2e/helm-charts
+terraform/eks/daemon/credentials/pod_identity/helm-charts
+terraform/eks/daemon/ebs/helm-charts
+terraform/eks/daemon/entity/helm-charts

docs/resources/dummy-neuron-monitor/Dockerfile

@@ -27,6 +27,7 @@ RUN apt-get update \
  && rm -rf /tmp/tmp* \
  && apt-get clean
 
+COPY neuron-monitor-output.json /opt/aws/neuron/bin/neuron-monitor-output.json
 COPY dummy_neuron_monitor.py /opt/aws/neuron/bin/dummy_neuron_monitor.py
 RUN chmod 755 /opt/aws/neuron/bin/dummy_neuron_monitor.py
 RUN pip3 install prometheus_client boto3 requests

docs/resources/dummy-neuron-monitor/dummy_neuron_monitor.py

@@ -66,6 +66,8 @@ type MetaData struct {
 	SampleApp                                   string
 	AccountId                                   string
 	EKSInstallationType                         eksinstallationtype.EKSInstallationType
+	PerformanceMetricMapName                    string
+	PerformanceTestName                         string
 }
 
 type MetaDataStrings struct {
@@ -110,6 +112,8 @@ type MetaDataStrings struct {
 	SampleApp                                   string
 	AccountId                                   string
 	EKSInstallationType                         string
+	PerformanceMetricMapName                    string
+	PerformanceTestName                         string
 }
 
 func registerComputeType(dataString *MetaDataStrings) {
@@ -140,6 +144,8 @@ func registerEKSData(d *MetaDataStrings) {
 	flag.StringVar(&(d.EKSClusterName), "eksClusterName", "", "EKS cluster name")
 	flag.StringVar(&(d.EksDeploymentStrategy), "eksDeploymentStrategy", "", "Daemon/Replica/Sidecar")
 	flag.StringVar(&(d.EksGpuType), "eksGpuType", "", "nvidia/inferentia")
+	flag.StringVar(&(d.PerformanceMetricMapName), "performanceMetricMapName", "", "name of eks performance metric map within directory")
+	flag.StringVar(&(d.PerformanceTestName), "performanceTestName", "", "name of eks performance test")
 }
 
 func registerEKSE2ETestData(dataString *MetaDataStrings) {
@@ -282,6 +288,8 @@ func fillEKSData(e *MetaData, data *MetaDataStrings) {
 
 	e.EKSClusterName = data.EKSClusterName
 	e.EksGpuType = data.EksGpuType
+	e.PerformanceMetricMapName = data.PerformanceMetricMapName
+	e.PerformanceTestName = data.PerformanceTestName
 }
 
 func fillEKSInstallationType(e *MetaData, data *MetaDataStrings) {

docs/resources/dummy-neuron-monitor/neuron-monitor-output.json

@@ -220,6 +220,8 @@ var testTypeToTestConfig = map[string][]testConfig{
 		{testDir: "../../../test/restart"},
 		{testDir: "../../../test/acceptance"},
 		{testDir: "../../../test/feature/windows/event_logs"},
+		{testDir: "../../../test/feature/windows/eventid_logs"},
+		{testDir: "../../../test/feature/windows/event_regex_logs"},
 		{testDir: "../../../test/log_state/logfile"},
 		{testDir: "../../../test/log_state/windows_event_log"},
 		{
@@ -339,6 +341,11 @@ var testTypeToTestConfig = map[string][]testConfig{
 			testDir: "./test/metric_value_benchmark", terraformDir: "terraform/eks/daemon/credentials/pod_identity",
 			targets: map[string]map[string]struct{}{"arc": {"amd64": {}}},
 		},
+		{
+			testDir:      "./test/ebscsi",
+			terraformDir: "terraform/eks/daemon/ebs",
+			targets: map[string]map[string]struct{}{"arc": {"amd64": {}}},
+		},
 	},
 	"eks_deployment": {
 		{testDir: "./test/metric_value_benchmark"},

environment/metadata.go

@@ -152,6 +152,8 @@ resource "null_resource" "validator" {
       -ecsLaunchType=FARGATE \
       -ecsDeploymentStrategy=DAEMON \
       -clusterArn=${aws_ecs_cluster.cluster.arn} \
+      -cwagentECSServiceName=${aws_ecs_service.cwagent_service.name} \
+      -cwagentConfigSsmParamName=${aws_ssm_parameter.cwagent_config.name} \
       -v
     EOT
   }

generator/test_case_generator.go

@@ -395,7 +395,7 @@ resource "kubernetes_daemonset" "neuron_monitor" {
         }
         container {
           name  = "neuron-monitor-prometheus"
-          image = "506463145083.dkr.ecr.us-west-2.amazonaws.com/mocked-neuron-monitor:v2"
+          image = "506463145083.dkr.ecr.us-west-2.amazonaws.com/mocked-neuron-monitor:v4"
           port {
             container_port = 8000
           }

terraform/ecs_fargate/linux/main.tf

@@ -194,17 +194,13 @@ resource "aws_security_group_rule" "nodes_cluster_inbound" {
   type                     = "ingress"
 }
 
-resource "null_resource" "clone_helm_chart" {
-  triggers = {
-    timestamp = "${timestamp()}" # Forces re-run on every apply
-  }
-  provisioner "local-exec" {
-    command = <<-EOT
-      if [ ! -d "./helm-charts" ]; then
-        git clone -b ${var.helm_chart_branch} https://github.com/aws-observability/helm-charts.git ./helm-charts
-      fi
-    EOT
-  }
+data "external" "clone_helm_chart" {
+  program = ["bash", "-c", <<-EOT
+    rm -rf ./helm-charts
+    git clone -b ${var.helm_chart_branch} https://github.com/aws-observability/helm-charts.git ./helm-charts
+    echo '{"status":"ready"}'
+  EOT
+  ]
 }
 
 resource "helm_release" "aws_observability" {
@@ -213,21 +209,22 @@ resource "helm_release" "aws_observability" {
   namespace        = "amazon-cloudwatch"
   create_namespace = true
 
-  set {
-    name  = "clusterName"
-    value = aws_eks_cluster.this.name
-  }
-
-  set {
-    name  = "region"
-    value = "us-west-2"
-  }
+  set = [
+    {
+      name  = "clusterName"
+      value = aws_eks_cluster.this.name
+    },
+    {
+      name  = "region"
+      value = var.region
+    }
+  ]
   depends_on = [
     aws_eks_cluster.this,
     aws_eks_node_group.this,
-    null_resource.clone_helm_chart,
     aws_eks_addon.pod_identity_addon,
     aws_eks_pod_identity_association.association,
+    data.external.clone_helm_chart,
   ]
 }
 

terraform/eks/daemon/awsneuron/main.tf

@@ -17,10 +17,10 @@ provider "kubernetes" {
 }
 
 provider "helm" {
-  kubernetes {
+  kubernetes = {
     host                   = aws_eks_cluster.this.endpoint
     cluster_ca_certificate = base64decode(aws_eks_cluster.this.certificate_authority.0.data)
-    exec {
+    exec = {
       api_version = "client.authentication.k8s.io/v1beta1"
       args        = ["eks", "get-token", "--cluster-name", aws_eks_cluster.this.name]
       command     = "aws"