[Microsoft] Run Azure Pipelines (VSTS) Agent on Docker or Azure Container Instance

Azure Pipelines Run the agent with Docker.

What is Azure Pipelines?

A service that builds / deploys like Jenkins.

What is Azure Pipelines Agent?

It is a daemon that actually operates the build and deploy process in response to instructions from Azure Pipelines.

Azure Pipelines comes with a free version, but there is a monthly run time limit. By preparing your own agent, you can use it without the maximum time.

What you need in advance

Organization name

See the URL for Azure DevOps. I think it looks like https://dev.azure.com/hogehoge. This is the part of hogehoge.

Personal Access Token

This is what the agent needs to connect to Azure DevOps.

Click here for [How to get Personal Access Token](https://docs.microsoft.com/ja-jp/azure/devops/organizations/accounts/use-personal-access-tokens-to-authenticate?view=azure-devops&tabs = preview-page).

Set the following as the scope.

• Agent Pools
  • Read & manage
• Code
  • Read
• Connected server
  • Connected server
• Deployment Groupos
  • Read & manage

Run with Docker

The official image of Micorosoft is Deprecated, so I created it though I'm afraid. sengokyu/azure-pipelines-agent:latest Pull as usual.

docker pull sengokyu/azure-pipelines-agent

At the time of execution, give the organization name and Personal Access Token prepared in advance as environment variables.

docker run -d -e TOKEN=${PAT} -e ORG=${ORG} sengokyu/azure-pipelines-agent

Run on Azure Container Instance

Deploy using an ARM template.

The resource definition part of the ARM template. The whole thing is on GitHub.

template.json

  "resources": [
    {
      "name": "[parameters('containerGroupName')]",
      "type": "Microsoft.ContainerInstance/containerGroups",
      "apiVersion": "2019-12-01",
      "location": "[resourceGroup().location]",
      "properties": {
        "containers": [
          {
            "name": "[parameters('agent')]",
            "properties": {
              "image": "[variables('containerImage')]",
              "resources": {
                "requests": {
                  "cpu": "[parameters('cpu')]",
                  "memoryInGb": "[parameters('memoryInGb')]"
                }
              },
              "environmentVariables": [
                { "name": "TOKEN", "secureValue": "[parameters('token')]" },
                { "name": "ORG", "value": "[parameters('org')]" },
                { "name": "POOL", "value": "[parameters('pool')]" },
                { "name": "AGENT", "value": "[parameters('agent')]" }
              ]
            }
          }
        ],
        "osType": "Linux",
        "restartPolicy": "Never"
      }
    }
  ]

Create a file that describes each parameter.

parameters.json

{
  "$schema": "https://schema.management.azure.com/schemas/2015-01-01/deploymentParameters.json#",
  "contentVersion": "1.0.0.0",
  "parameters": {
    "containerGroupName": { "value": "vsts-agent" },
    "pool": { "value": "Default" },
    "agent": { "value": "vsts-agent-on-aci" },
    "token": { "value": "YourPersonalAccessToken" },
    "org": { "value": "YourOrganizationName" },
    "cpu": { "value": 4 },
    "memoryInGb": { "value": 8 }
  }
}

Execute the az command.

az deployment group create -f template.json -p @parameters.json -g ResourceGroupName