{"id":368,"date":"2026-01-23T07:06:28","date_gmt":"2026-01-23T07:06:28","guid":{"rendered":"https:\/\/blog.ngocha.biz\/?p=368"},"modified":"2026-01-23T07:06:28","modified_gmt":"2026-01-23T07:06:28","slug":"setup-self-hosted-runner-azure-devops-pipeline","status":"publish","type":"post","link":"https:\/\/blog.ngocha.biz\/?p=368","title":{"rendered":"How to Setup Self Hosted Linux Runner for Azure DevOps?"},"content":{"rendered":"

In this blog, you will learn how to set up a self-hosted linux based runner<\/strong> for Azure DevOps<\/strong> pipeline.<\/p>\n

By the end of this blog, you will learn<\/p>\n

    \n
  1. What is a self-hosted runner?<\/li>\n
  2. How to set up a self-hosted Linux runner in Azure DevOps?<\/li>\n
  3. How to test self hosted runner in Azure DevOps?<\/li>\n
  4. Best practices for self-hosted runners.<\/li>\n<\/ol>\n

    Before we begin the setup, lets understand what is a self-hosted runner in Azure.<\/p>\n

    What is a Self-Hosted Runner ?<\/h2>\n

    Self-hosted runners<\/a> are servers or containers that we manage to run our Azure DevOps pipeline tasks.<\/p>\n

    Self-hosted<\/strong> means we control the environment where the pipeline runners execute. This setup is typically used for production-level projects<\/strong>, where we need more control over security, performance, and dependencies.<\/p>\n

    The following can be a self-hosted runner.<\/p>\n

      \n
    1. Virtual machine<\/li>\n
    2. Physical server<\/li>\n
    3. Container<\/a><\/li>\n
    4. Kubernetes<\/a><\/li>\n<\/ol>\n

      Organizations generally use self-hosted runners<\/strong> because they can control performance, cost, security, networking, compliance, and other aspects.<\/p>\n

      By having control over the runners, you can install all required security tools, logging agents, build tools, and packages that are needed to meet the organization\u2019s compliance requirements.<\/strong><\/p>\n

      \n
      \ud83d\udca1<\/div>\n
      For production use cases, it is recommended to use self-hosted runners inside the organization\u2019s network. This helps avoid data exposure and allows secure access to internal resources.<\/div>\n<\/div>\n

      Setup Prerequisites<\/h2>\n

      The following are the requirements to setup a self hosted runner.<\/p>\n

        \n
      1. Azure DevOps account with required permissions<\/li>\n
      2. Azure DevOps project with a pipeline.<\/li>\n
      3. A virtual machine (e.g., AWS EC2<\/a>, Azure VM, etc)<\/li>\n<\/ol>\n
        \n
        \ud83d\udccc<\/div>\n
        In this setup we are using an Azure Ubuntu 24.04 VM as the runner.<\/div>\n<\/div>\n

        Setting Up an Azure DevOps Self-Hosted Runner<\/h2>\n

        Here, we are going to use a virtual machine as a runner.<\/p>\n

        Assuming you already have an Azure VM ready.<\/p>\n

        Now, we can start with the Personal Token creation.<\/p>\n

        Step 1: Generate a Personal Access Token (PAT)<\/h3>\n

        The runner needs a Personal Access Token (PAT) to authenticate with Azure DevOps.<\/p>\n

        To create a PAT, open Azure DevOps<\/strong> –> login to the Azure DevOps organization<\/strong> –> Select Project<\/strong>.<\/p>\n

        On the project console, click the user icon on the right side corner and select the “User settings<\/strong>“<\/p>\n

        \"selecting<\/figure>\n

        Select the “Personal access tokens<\/strong>” option to create a new one.<\/p>\n

        \"selecting<\/figure>\n

        Click the “+ New Token<\/strong>” button.<\/p>\n

        \"creating<\/figure>\n

        On the creation section, you need to provide a name for the PAT, select the organization, expiration date, and the permission level.<\/p>\n

        \n
        \ud83d\udca1<\/div>\n
        Here, in the scope, I am giving full access, but in real projects, select “Custom defined<\/strong><\/b>” and only give the least privileges.<\/div>\n<\/div>\n
        \"setting<\/figure>\n

        Once the creation is completed, you will see a success message and have a warning to save the generated PAT.<\/p>\n

        \"the<\/figure>\n

        By default, the PAT will not be stored anywhere, so before closing the window, store the PAT.<\/p>\n

        In the next section, we will configure the runner.<\/p>\n

        Step 2: Create an Agent Pool on the Azure DevOps Project.<\/h3>\n

        To configure the project-level runner, go to project settings<\/strong> and select the “Agent pools” then click the “Add pool” button.<\/p>\n

        \"the<\/figure>\n

        On the agent pool form, select the pool link as “New” and select the pool type as “Self-hosted”, give a name, and select the permission, then click the “Create” button to create the new pool.<\/p>\n

        \"filling<\/figure>\n
        \n
        \ud83d\udca1<\/div>\n
        If you are not sure about the pipeline workload, you can choose the “Azure virtual machine scale set<\/strong><\/b>“. So that the number of VMs will automatically scale up and down based on the workload.<\/div>\n<\/div>\n

        Select the created pool to add the agents.<\/p>\n

        \"selecting<\/figure>\n
        \n
        \ud83d\udca1<\/div>\n
        We can even create an organization-level runner in Azure DevOps from the Organization settings.<\/p>\n

        These runners can be used for all projects across the organization.<\/p><\/div>\n<\/div>\n

        Step 3: Create Agent in Azure DevOps Project<\/h3>\n

        Once open the pool, click the “New agent<\/strong>” to start the configuration.<\/p>\n

        \"creating<\/figure>\n

        The next page will give you the instructions and commands to run on the server to turn it into a runner.<\/p>\n

        \"the<\/figure>\n

        Here, I have selected the Linux OS because, we are going to use the Ubuntu server.<\/p>\n

        Before we execute the commands on the virtual machine, we need the following prerequisites.<\/p>\n

          \n
        1. Server URL – E.g., https:\/\/dev.azure.com\/<ORGANIZATION NAME><\/code><\/li>\n
        2. PAT<\/li>\n
        3. Agent Pool name<\/li>\n<\/ol>\n

          Now, SSH<\/a> to the virtual Machine.<\/p>\n

          Create a directory<\/p>\n

          mkdir myagent && cd myagent<\/code><\/pre>\n
          Download the runner setup package<\/p>\n
          wget https:\/\/download.agent.dev.azure.com\/agent\/4.261.0\/vsts-agent-linux-x64-4.261.0.tar.gz<\/code><\/pre>\n
          Extract the package<\/p>\n
          tar zxvf vsts-agent-linux-x64-4.261.0.tar.gz<\/code><\/pre>\n
          Configure the agent<\/p>\n
          .\/config.sh<\/code><\/pre>\n
          Once you reach the .\/config<\/code> step, it will prompt you to provide the following.<\/p>\n
            \n
          1. License agreement – Type y<\/code><\/li>\n
          2. Server URL – https:\/\/dev.azure.com\/<ORG NAME><\/code><\/li>\n
          3. Authentication type – Press Enter<\/code> for PAT and paste the PAT<\/li>\n
          4. Agent pool name <\/li>\n
          5. Name for the Agent<\/li>\n<\/ol>\n
            azureuser@azure-runner-vm:~\/myagent$ .\/config.sh\n\n  ___                      ______ _            _ _\n \/ _ \\                     | ___ (_)          | (_)\n\/ \/_\\ \\_____   _ _ __ ___  | |_\/ \/_ _ __   ___| |_ _ __   ___  ___\n|  _  |_  \/ | | | '__\/ _ \\ |  __\/| | '_ \\ \/ _ \\ | | '_ \\ \/ _ \\\/ __|\n| | | |\/ \/| |_| | | |  __\/ | |   | | |_) |  __\/ | | | | |  __\/\\__ \\\n\\_| |_\/___|\\__,_|_|  \\___| \\_|   |_| .__\/ \\___|_|_|_| |_|\\___||___\/\n                                   | |\n        agent v4.261.0             |_|          (commit 45f3f01)\n\n\n>> End User License Agreements:\n\nBuilding sources from a TFVC repository requires accepting the Team Explorer Everywhere End User License Agreement. This step is not required for building sources from Git repositories.\n\nA copy of the Team Explorer Everywhere license agreement can be found at:\n  \/home\/azureuser\/myagent\/license.html\n\nEnter (Y\/N) Accept the Team Explorer Everywhere license agreement now? (press enter for N) > y\n\n>> Connect:\n\nEnter server URL > https:\/\/dev.azure.com\/devopscube-org\nEnter authentication type (press enter for PAT) > \nEnter personal access token > ************************************************************************************\nConnecting to server ...\n\n>> Register Agent:\n\nEnter agent pool (press enter for default) > custom-linux-pool\nEnter agent name (press enter for azure-runner-vm) > linux-agent\nEnter replace? (Y\/N) (press enter for N) > n\nScanning for tool capabilities.\nConnecting to the server.\nSuccessfully added the agent\nTesting agent connection.\nEnter work folder (press enter for _work) > \n2025-10-22 09:51:04Z: Settings Saved.<\/code><\/pre>\n
            The configurations are ready now, so we need to make them run the agent.<\/p>\n
            .\/run.sh<\/code><\/pre>\n
            You will see the following output.<\/p>\n
            $ .\/run.sh\n\nScanning for tool capabilities.\nConnecting to the server.\n2025-10-22 09:58:14Z: Listening for Jobs<\/code><\/pre>\n
            This output ensures that the agent is configured and running. We can confirm this from the dashboard.<\/p>\n
            \"the<\/figure>\n
            The console output also shows that the runner is active. So we can test now.<\/p>\n
            Testing the Self-Hosted Azure DevOps Runner<\/h2>\n
            To test the self-hosted runner, we need to call it from a pipeline.<\/p>\n
            We create a fresh test pipeline. For that, navigate to Pipelines –> New pipeline –> Azure Repos Git –> [Select Repository] –> Starter pipeline<\/strong>.<\/p>\n
            Now, you can see a YAML file with some default configurations. Replace the entire contents with the following.<\/p>\n
            trigger:\n- main\n\npool:\n  name: 'custom-linux-pool'\n  demands:\n  - Agent.Name -equals linux-agent\n\nsteps:\n- script: whoami\n  displayName: 'Check current user'\n\n<\/code><\/pre>\n
            Here,<\/p>\n