Skip to content

Latest commit

 

History

History
455 lines (424 loc) · 18.9 KB

SETUP_GUIDE.md

File metadata and controls

455 lines (424 loc) · 18.9 KB

Enabling Application Signals for .NET with OpenTelemetry Auto-Instrumentation

Introduction

This guide provides step-by-step instructions for enabling Application Signals in .NET applications using the AWS Distro for OpenTelemetry (ADOT) Instrumentation for .NET on supported platforms.

  • The Integration Test Application is used as the reference .NET application throughout this guide. to demonstrates how to implement and test the ADOT instrumentation
  • The CloudWatch Agent is set up to collect OTel telemetry data emitted by the application, and then sent to Amazon CloudWatch for monitoring and analysis
  • The guide provisions all necessary infrastructure to run the sample application from scratch

EC2 Linux

Prerequisites

  • Launch a Linux EC2 instance
  • Log into the instance using SSH client
    chmod 400 "<customer_key_pair_name>.pem"
    ssh -i "<customer_key_pair_name>.pem" [email protected]
  • Install .NET 8
    sudo rpm --import https://packages.microsoft.com/keys/microsoft.asc
    sudo wget -O /etc/yum.repos.d/microsoft-prod.repo https://packages.microsoft.com/config/fedora/37/prod.repo
    sudo dnf install -y dotnet-sdk-8.0
    dotnet --version > /tmp/dotnet-version
  • Install AWS CLI
    • AWS CLI is preinstalled on Amazon Linux instances, verify by
      aws --version
    • If not installed, follow AWS CLI Installation Guide to install
  • Install Git
    sudo yum update -y
    sudo yum install git -y
    git --version

Set up CloudWatch Agent with Application Signals enabled

Download AWS Distro of OTel .NET auto-instrumentation agent

wget https://github.com/aws-observability/aws-otel-dotnet-instrumentation/releases/latest/download/aws-distro-opentelemetry-dotnet-instrumentation-linux-glibc-x64.zip
export INSTALL_DIR=~/OpenTelemetryDistribution
unzip AWS-opentelemetry-dotnet-instrumentation-linux-glibc-x64.zip -d $INSTALL_DIR

Run the sample app with .NET auto-instrumentation agent

  • Pull the agent repo from GitHub to use the integration test application
    git clone https://github.com/aws-observability/aws-otel-dotnet-instrumentation
  • Build the sample application
    cd aws-otel-dotnet-instrumentation/sample-applications/integration-test-app/
    dotnet publish integration-test-app/integration-test-app.csproj -c Release -o out
  • Run the sample applcation with .NET auto-instrumentation agent
    export INSTALL_DIR=~/OpenTelemetryDistribution
    export CORECLR_ENABLE_PROFILING=1
    export CORECLR_PROFILER={918728DD-259F-4A6A-AC2B-B85E1B658318}
    export CORECLR_PROFILER_PATH=${INSTALL_DIR}/linux-x64/OpenTelemetry.AutoInstrumentation.Native.so
    export DOTNET_ADDITIONAL_DEPS=${INSTALL_DIR}/AdditionalDeps
    export DOTNET_SHARED_STORE=${INSTALL_DIR}/store
    export DOTNET_STARTUP_HOOKS=${INSTALL_DIR}/net/OpenTelemetry.AutoInstrumentation.StartupHook.dll
    export OTEL_DOTNET_AUTO_HOME=${INSTALL_DIR}
    
    export OTEL_DOTNET_AUTO_PLUGINS="AWS.Distro.OpenTelemetry.AutoInstrumentation.Plugin, AWS.Distro.OpenTelemetry.AutoInstrumentation"
    
    export OTEL_RESOURCE_ATTRIBUTES=service.name=aws-otel-integ-test
    export OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf
    export OTEL_EXPORTER_OTLP_ENDPOINT=http://127.0.0.1:4316
    export OTEL_AWS_APPLICATION_SIGNALS_EXPORTER_ENDPOINT=http://127.0.0.1:4316/v1/metrics
    export OTEL_METRICS_EXPORTER=none
    export OTEL_AWS_APPLICATION_SIGNALS_ENABLED=true
    export OTEL_TRACES_SAMPLER=xray
    export OTEL_TRACES_SAMPLER_ARG=http://127.0.0.1:2000
    
    
    export ASPNETCORE_URLS=http://+:8080
    export LISTEN_ADDRESS=0.0.0.0:8080
    
    dotnet out/integration-test-app.dll

View the collected OTel telemetry data in the CloudWatch Application Signals service console

EC2 Windows

Prerequisites

  • Launch a Windows EC2 instance
  • Log into the instance using RDP tools
    • Open a PowerShell session
  • Install .NET 8
    • In powershell session
      wget -O dotnet-install.ps1 https://dot.net/v1/dotnet-install.ps1
      .\dotnet-install.ps1 -Version 8.0.302
  • Install AWS CLI
    • Install following AWS CLI user guide
      msiexec.exe /i https://awscli.amazonaws.com/AWSCLIV2.msi
    • Reopen PowerShell session after around 30 seconds and verify installation
      aws --version
  • Install Git
    winget install --id Git.Git -e --source winget
    $env:PATH += ";C:\Program Files\Git\cmd"
    git --version

Set up CloudWatch Agent with Application Signals enabled

  • Enable Application Signals in your account
  • Download and start the CloudWatch agent
    • Either follow Linux instruction and refer to CW Agent guidance for different processes setting up Windows
    • Or follow these steps:
      • Install CloudWatch Agent
      • Create CloudWatch Agent configuraiton file
        New-Item -ItemType Directory -Path "C:\tmp" -Force
        
        $content = @"
        {
          "traces": {
            "traces_collected": {
              "application_signals": {}
            }
          },
          "logs": {
            "metrics_collected": {
              "application_signals": {}
            }
          }
        }
        "@
        
        Set-Content -Path "C:\tmp\application-signals-cwagent-config.txt" -Value $content
      • Start CloudWatch Agent
        & $Env:ProgramFiles\Amazon\AmazonCloudWatchAgent\amazon-cloudwatch-agent-ctl.ps1 -m ec2 -a status

Download AWS Distro of OTel .NET auto-instrumentation agent

Invoke-WebRequest -Uri "https://github.com/aws-observability/aws-otel-dotnet-instrumentation/releases/latest/download/aws-distro-opentelemetry-dotnet-instrumentation-windows.zip" -OutFile "aws-distro-opentelemetry-dotnet-instrumentation-windows.zip"
$env:INSTALL_DIR = "C:\Users\Administrator\Downloads\OpenTelemetryDistribution"
Expand-Archive -Path "aws-distro-opentelemetry-dotnet-instrumentation-windows.zip" -DestinationPath $env:INSTALL_DIR

Run the sample app with .NET auto-instrumentation agent

  • Pull the agent repo from GitHub to use the integration test application
    git clone https://github.com/aws-observability/aws-otel-dotnet-instrumentation
  • Build the sample application
    cd aws-otel-dotnet-instrumentation/sample-applications/integration-test-app/
    dotnet publish integration-test-app/integration-test-app.csproj -c Release -o out
  • Run the sample applcation with .NET auto-instrumentation agent
    $env:INSTALL_DIR = "C:\Users\Administrator\Downloads\OpenTelemetryDistribution"
    $env:CORECLR_ENABLE_PROFILING = "1"
    $env:CORECLR_PROFILER = "{918728DD-259F-4A6A-AC2B-B85E1B658318}"
    $env:CORECLR_PROFILER_PATH = "$env:INSTALL_DIR\win-x64\OpenTelemetry.AutoInstrumentation.Native.dll"
    $env:DOTNET_ADDITIONAL_DEPS = "$env:INSTALL_DIR\AdditionalDeps"
    $env:DOTNET_SHARED_STORE = "$env:INSTALL_DIR\store"
    $env:DOTNET_STARTUP_HOOKS = "$env:INSTALL_DIR\net\OpenTelemetry.AutoInstrumentation.StartupHook.dll"
    $env:OTEL_DOTNET_AUTO_HOME = "$env:INSTALL_DIR"
    $env:OTEL_DOTNET_AUTO_PLUGINS = "AWS.Distro.OpenTelemetry.AutoInstrumentation.Plugin, AWS.Distro.OpenTelemetry.AutoInstrumentation"
    $env:OTEL_RESOURCE_ATTRIBUTES = "service.name=aws-otel-integ-test"
    $env:OTEL_EXPORTER_OTLP_PROTOCOL = "http/protobuf"
    $env:OTEL_EXPORTER_OTLP_ENDPOINT = "http://127.0.0.1:4316"
    $env:OTEL_AWS_APPLICATION_SIGNALS_EXPORTER_ENDPOINT = "http://127.0.0.1:4316/v1/metrics“
    $env:OTEL_METRICS_EXPORTER = "none"
    $env:OTEL_AWS_APPLICATION_SIGNALS_ENABLED = "true"
    $env:OTEL_TRACES_SAMPLER = "xray"
    $env:OTEL_TRACES_SAMPLER_ARG = "http://127.0.0.1:2000"
    $env:ASPNETCORE_URLS = "http://+:8080"
    $env:LISTEN_ADDRESS = "0.0.0.0:8080"
    
    dotnet out/integration-test-app.dll

View the collected OTel telemetry data in the CloudWatch Application Signals service console

EKS Linux

Prerequisites

  • Create VPC for EKS cluster
    • Choose IPv4 CloudFormation template for easier setup
  • Install kubectl
    curl -O https://s3.us-west-2.amazonaws.com/amazon-eks/1.30.0/2024-05-12/bin/darwin/amd64/kubectl
  • Install AWS CLI
  • Install eksctl
    brew tap weaveworks/tap
    brew install weaveworks/tap/eksctl
  • Create EKS Cluster
    • Specify --vpc-private-subnets to be the private subnets created in the VPC CloudFormation stack
    eksctl create cluster --name <my-cluster-name> --region <region-code> --version 1.30 --vpc-private-subnets subnet-ID1,subnet-ID2 --without-nodegroup

Launch EKS Worker Nodes

  • Launch Linux Workder Nodes
    • Both the sample application and amazon-cloudwatch-observability addon run on pods on these Linux nodes
    • Specify --subnet-ids to be the public subnets created in the VPC CloudFormation stack
      eksctl create nodegroup --cluster <my-cluster-name> --name <my-nodegroup-name> --node-type m5.large --nodes 3 --nodes-min 1 --nodes-max 4 --ssh-access --managed --region <region-code> --ssh-public-key <my-public-key-name> --subnet-ids subnet-ID1,subnet-ID2
    • Add managed policy CloudWatchAgentServerPolicy to the node role

Install amazon-cloudwatch-observability addon

  • Install addon
    aws eks create-addon --cluster-name <my-cluster-name> --addon-name amazon-cloudwatch-observability
  • Verify deployment configuration of the addon
    kubectl edit deployment amazon-cloudwatch-observability-controller-manager -n amazon-cloudwatch
    • Verify ADOT auto-instrumentation image is specified
      spec:
        containers:
        - args:
          ....
          - --auto-instrumentation-dotnet-image=public.ecr.aws/aws-observability/adot-autoinstrumentation-dotnet:v1.1.0

Build ECR image of sample application

  • Pull the agent repo from GitHub to use the integration test application
    git clone https://github.com/aws-observability/aws-otel-dotnet-instrumentation
  • Build the sample application
    cd aws-otel-dotnet-instrumentation/sample-applications/integration-test-app/
    dotnet publish integration-test-app/integration-test-app.csproj -c Release -o out
  • Replace Dockerfile contents
    FROM mcr.microsoft.com/dotnet/sdk:8.0 AS build-env
    WORKDIR /app
    COPY . ./
    RUN dotnet publish integration-test-app/integration-test-app.csproj -c Release -o out
    
    FROM mcr.microsoft.com/dotnet/aspnet:8.0
    WORKDIR /app
    EXPOSE 8080
    COPY --from=build-env /app/out .
    ENTRYPOINT ["dotnet", "integration-test-app.dll"]
  • Build docker image for sample applcation
    docker build -t dotnet-sample-app .
    • Verify sample-app docker image created
      docker images dotnet-sample-app
  • Authenticate docker to your Amazon ECR Registry
    aws ecr get-login-password --region <region-code> | docker login --username AWS --password-stdin <aws-account-id>.dkr.ecr.<region-code>.amazonaws.com
  • Create a repository in Amazon ECR
    aws ecr create-repository --repository-name dotnet-sample-app-repo --region <region-code>
  • Tag your local docker image with the ECR repository URI
    docker tag dotnet-sample-app:latest <aws-account-id>.dkr.ecr.<region-code>.amazonaws.com/dotnet-sample-app-repo:latest
  • Push the docker image to your ECR repository
    docker push <aws-account-id>.dkr.ecr.<region-code>.amazonaws.com/dotnet-sample-app-repo:latest

Run the sample app with .NET auto-instrumentation agent

  • Create sample application configuration file dotnet-demo.yaml
    apiVersion: v1
    kind: Namespace
    metadata:
      labels:
        kubernetes.io/metadata.name: linux
      name: dotnet
    ---
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: dotnet-demo
      namespace: dotnet
    spec:
      selector:
        matchLabels:
          app: dotnet-demo
          tier: backend
          track: stable
      replicas: 1
      template:
        metadata:
          labels:
            app: dotnet-demo
            tier: backend
            track: stable
          annotations:
            instrumentation.opentelemetry.io/inject-dotnet: "true"
        spec:
          containers:
          - name: dotnet-demo
            image: <aws-account-id>.dkr.ecr.<region-code>.amazonaws.com/dotnet-sample-app-repo:latest
            ports:
            - name: http
              containerPort: 8080
            command: ["dotnet", "integration-test-app.dll"]
            imagePullPolicy: Always
            env:
              - name: AWS_REGION
                value: "us-east-1"
              - name: LISTEN_ADDRESS
                value: "0.0.0.0:8080"
              - name: ASPNETCORE_URLS
                value: "http://+:8080"
          nodeSelector:
            kubernetes.io/os: linux
          tolerations:
              - key: "os"
                operator: "Equal"
                value: "linux"
                effect: "NoSchedule"
  • Run the sample applcation with .NET auto-instrumentation agent
    kubectl apply -f dotnet-demo.yaml

View the collected OTel telemetry data in the CloudWatch Application Signals service console

EKS Windows

Prerequisites

Launch EKS Worker Nodes

  • Launch Linux Workder Nodes
    • amazon-cloudwatch-observability addon runs on pods on these Linux nodes
    • Specify --subnet-ids to be the public subnets created in the VPC CloudFormation stack
      eksctl create nodegroup --cluster <my-cluster-name> --name <my-linux-nodegroup-name> --node-type m5.large --nodes 3 --nodes-min 1 --nodes-max 4 --ssh-access --managed --region <region-code> --ssh-public-key <my-public-key-name> --subnet-ids subnet-ID1,subnet-ID2
    • Add managed policy CloudWatchAgentServerPolicy to the node role
  • Launch Windows Workder Nodes
    • The sample application runs on pods on these Windows nodes
    • Specify --subnet-ids to be the public subnets created in the VPC CloudFormation stack
      eksctl create nodegroup --cluster <my-cluster-name> --name <my-windows-nodegroup-name> --node-type m5.large --nodes 3 --nodes-min 1 --nodes-max 4 --managed --region <region-code> --node-ami-family WindowsServer2022FullContainer --subnet-ids subnet-ID1,subnet-ID2
    • Follow EKS guidance to Enable Windows support
    • Open Windows Server Firewall rules
      • This is needed since Windows nodes using the CloudWatch Agent as a Host Process Container cannot enable Application Signals due to limitations in Kubernetes networking on Windows
      • Log into each Windows node (using RDP tool, or Systems Manager Session Manager) and run PowerShell command
        netsh advfirewall firewall add rule name="appsignals" dir=in action=allow localport=4316 protocol=tcp

Install amazon-cloudwatch-observability addon

Build ECR image of sample application

  • Same as EKS Linux, replace Dockerfile contents with
FROM mcr.microsoft.com/dotnet/aspnet:8.0-nanoserver-ltsc2022
WORKDIR /app
EXPOSE 8080
COPY out .
ENTRYPOINT ["dotnet", "integration-test-app.dll"]

Run the sample app with .NET auto-instrumentation agent

  • Create sample application configuration file dotnet-demo-windows.yaml
    apiVersion: v1
    kind: Namespace
    metadata:
      labels:
        kubernetes.io/metadata.name: windows
      name: dotnet-win
    ---
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: dotnet-demo-win
      namespace: dotnet-win
    spec:
      selector:
        matchLabels:
          app: dotnet-demo-win
          tier: backend
          track: stable
      replicas: 1
      template:
        metadata:
          labels:
            app: dotnet-demo-win
            tier: backend
            track: stable
          annotations:
            instrumentation.opentelemetry.io/inject-dotnet: "true"
        spec:
          containers:
          - name: dotnet-demo-win
            image: <aws-account-id>.dkr.ecr.<region-code>.amazonaws.com/dotnet-sample-app-repo:latest
            ports:
            - name: http
              containerPort: 8080
            command: ["dotnet", "integration-test-app.dll"]
            imagePullPolicy: Always
            env:
              - name: AWS_REGION
                value: "us-east-1"
              - name: LISTEN_ADDRESS
                value: "0.0.0.0:8080"
              - name: ASPNETCORE_URLS
                value: "http://+:8080"
          nodeSelector:
            kubernetes.io/os: windows
          tolerations:
              - key: "os"
                operator: "Equal"
                value: "windows"
                effect: "NoSchedule"
    
  • Run the sample applcation with .NET auto-instrumentation agent
    kubectl apply -f dotnet-demo-windows.yaml

View the collected OTel telemetry data in the CloudWatch Application Signals service console