Installing the Cloud Monitoring agent on individual VMs

The Cloud Monitoring agent gathers system and application metrics from your VM instances and sends them to Monitoring. You can also configure the Monitoring agent to monitor third-party applications.

This guide explains how to install the Cloud Monitoring agent for Monitoring on Compute Engine and Amazon Elastic Compute Cloud (EC2) VMs. Using the agent is optional but recommended for Compute Engine VM instances. For these VMs, Monitoring can access some metrics from the VM's hypervisor without the agent, including CPU utilization, some disk traffic metrics, network traffic, and uptime. Using the agent is required for Amazon Elastic Compute Cloud (EC2) VMs.

On instances running Microsoft Windows, the agent records CPU utilization and memory, pagefile, and volume usage. If you are running IIS or SQL server, the agent collects metrics from those services by default.

Before you begin

To install the agent, ensure that you have the following:

  • A supported VM instance in a Google Cloud project or Amazon Web Services (AWS) account.

    • A minimum of 250 MiB of resident (RSS) memory is recommended to run the Monitoring agent.

    Also ensure your VM is running a supported operating system.

  • Credentials on the VM instance that authorize communication with Cloud Logging or Cloud Monitoring. Compute Engine VM instances generally have the correct credentials by default. If either of the following scenarios applies to you, then you might not have the proper credentials and must complete the Authorize the Monitoring agent procedures:

    • Running AWS EC2 VM instances, you must install authorization credentials on your VMs before installing the agent.

    • Running very old Compute Engine instances or Compute Engine instances created without the default credentials.

    To check if you have the proper credentials, run the Verifying Compute Engine credentials procedures.

  • For AWS users, do the following:

    1. Connect your AWS account to a Google Cloud. For information about this process, see Collect metrics from AWS accounts.

    2. Refer to the Google Cloud projects for AWS EC2 VM instances section for additional information.

  • For pricing information, go to Pricing for Google Cloud Observability.

  • If you're using VMs that don't have access to remote package repositories, refer to the VMs without remote package access section for more information.

Installing the agent from the command line

To install the agent using the command line, use the following instructions.

Installing the latest version of the agent

To install the latest version of the agent, complete the following steps.

Linux

  1. Open a terminal connection to your VM instance using SSH or a similar tool and ensure you have sudo access.

  2. Change to a directory you have write access to, for example your home directory.

  3. Download and run the agent-installation script by using the following commands:

    curl -sSO https://dl.google.com/cloudagents/add-monitoring-agent-repo.sh
    sudo bash add-monitoring-agent-repo.sh --also-install
    

    After it is installed, the agent is started automatically.

Windows

  1. Connect to your instance using RDP or a similar tool and login to Windows.

  2. Open a PowerShell terminal with administrator privileges by right-clicking the PowerShell icon and selecting Run as Administrator.

  3. Run the following PowerShell commands:

    (New-Object Net.WebClient).DownloadFile("https://repo.stackdriver.com/windows/StackdriverMonitoring-GCM-46.exe", "${env:UserProfile}\StackdriverMonitoring-GCM-46.exe")
    & "${env:UserProfile}\StackdriverMonitoring-GCM-46.exe"
    

Installing a specific version of the agent

To install a specific version of the agent, complete the following steps.

Linux

  1. Open a terminal connection to your VM instance using SSH or a similar tool and ensure you have sudo access.

  2. Change to a directory you have write access to, for example your home directory.

  3. Download the agent installation script:

    curl -sSO https://dl.google.com/cloudagents/add-monitoring-agent-repo.sh
    

    When running the add-monitoring-agent-repo.sh script, you can also set the following flags:

    • --verbose: Turns on verbose logging during the script execution.
    • --also-install: Installs the agent after adding the agent package repository.
    • --version: Sets the agent version for the script to install.
    • --uninstall: Uninstalls the agent.
    • --remove-repo: Removes the corresponding agent package repository after installing or uninstalling the agent.
    • --dry-run: Triggers only a dry run of the script execution and prints out the commands that it is supposed to execute.

    See the script comments for more information and example usage.

  4. Add the agent's package repository and install the agent:

    1. To list the available agent versions in order to select which version to install, refer to Listing all agent versions.

    2. For production environments, you might want to pin to a major version to avoid installing major versions that might include backward incompatible changes. To pin to a major version, run:

      sudo bash add-monitoring-agent-repo.sh --also-install \
        --version=MAJOR_VERSION.*.*
      

      For example, to pin to the 6.x.x of the agent, run:

      sudo bash add-monitoring-agent-repo.sh --also-install \
        --version=6.*.*
      
    3. To install a specific version of the agent, run:

      sudo bash add-monitoring-agent-repo.sh --also-install \
        --version=MAJOR_VERSION.MINOR_VERSION.PATCH_VERSION
      
  5. Start the agent service

    sudo service stackdriver-agent start
    

You can delete the installation script after it runs successfully.

  • To verify that the agent is working as expected, run:

    sudo service stackdriver-agent status
    

    The status of the agent should be OK.

  • You can also examine the logs and ensure there are no errors:

    sudo grep collectd /var/log/{syslog,messages} | tail
    

If you have trouble with the installation, refer to the Troubleshooting page.

Windows

  1. Connect to your instance using RDP or a similar tool and login to Windows.

  2. Open a PowerShell terminal with administrator privileges by right-clicking the PowerShell icon and selecting Run as Administrator.

    1. Run the following PowerShell commands:

      (New-Object Net.WebClient).DownloadFile("https://repo.stackdriver.com/windows/StackdriverMonitoring-GCM-46.exe", "${env:UserProfile}\StackdriverMonitoring-GCM-46.exe")
      & "${env:UserProfile}\StackdriverMonitoring-GCM-46.exe"
      

    Alternatively, you can browse to the following URL to download and run the agent's installer:

    https://repo.stackdriver.com/windows/StackdriverMonitoring-GCM-46.exe

    To install the agent silently, append the /S option to the invocation of the installer:

    & "${env:UserProfile}\StackdriverMonitoring-GCM-46.exe" /S
    

    In “silent” mode use the /D option to specify the installation directory, for example:

    & "${env:UserProfile}\StackdriverMonitoring-GCM-46.exe" /S /D="C:\Stackdriver\Monitoring\"
    

    You can delete the installer when it completes successfully.

    If you have trouble with the installation, refer to the Troubleshooting page.

Viewing agent information by using the Google Cloud console

You can find status information about the agent on the pre-configured Monitoring VM Instances dashboard. To reach this dashboard, do the following:

In the navigation panel of the Google Cloud console, select Monitoring, select  Dashboards, and then select VM instances:

Go to VM Instances Dashboard

The List view on the Inventory tab on the dashboard lists all VMs and includes a status column for your agent, as shown in the following screenshot:

The VM Instances dashboard in Monitoring shows the status
of
agents.

The Agent column reports the following values:

  • Not detected: Either you don't have an agent installed or it is not running. If you aren't sure if you've installed an agent, then you can query for the installed version. If you've installed the agent, then you can restart the agent.

  • Ops Agent: You are running the Ops Agent. If you don't see a green checkmark beside the entry, then there is an agent upgrade available, based on the detected operating system of your VM.

    When you hover over the Ops Agent indicator in the table, you see information about the version of the Ops Agent. If you are running an older version, you also see a recommendation to upgrade your agent.

  • Pending: The Ops Agent is being installed or upgraded.

  • Legacy Agent: You are running the legacy Monitoring or the Logging agent. See Migrating from the legacy agents to the Ops Agent for information on transitioning to the Ops Agent.

  • Not applicable: This VM is not a supported platform for running the agent.

  • Unknown: The VM is not running, so the agent's status is not known.

You can install the Ops Agent by doing the following:

  • Select the VM instances on which you want to install agents.
  • Click the Install/Update Ops Agent option on the Instances table.

You can also install or update the Ops Agent from the VM Details page for a specific VM.

When possible, the agent is installed by using an Ops Agent OS policy. For more information, see Manage VMs covered by the Ops Agent OS policy. Ops Agent OS policies aren't supported on all versions of all operating systems. In this case, clicking Install/Update Ops Agent provides a series of commands to run in Cloud Shell.

The Ops Agent collects both metrics and logs by default. You can change this default behavior by configuring the Ops Agent.

Optional tasks

This section describes how to perform common maintenance tasks.

Configuring the Monitoring agent

To adjust the agent configuration, see Configure the Monitoring agent.

Configuring an HTTP proxy

If you use an HTTP proxy for proxying requests to the Logging and Monitoring APIs, do the following:

Linux

  1. Edit the following configuration file (create the file if it doesn't already exist):

    • For agent versions 6.0.0 and higher, edit:

      /etc/default/stackdriver-agent
      
    • For agent versions earlier than 6.0.0, edit the appropriate file for your OS:

      For Debian and Ubuntu, edit:

      /etc/default/stackdriver-agent
      

      For CentOS and SLES, edit:

      /etc/default/stackdriver-collectd
      
  2. Add the following to the file:

     export http_proxy="http://proxy-ip:proxy-port"
     export https_proxy="http://proxy-ip:proxy-port"
     export no_proxy=169.254.169.254  # Skip proxy for the local Metadata Server.
    
  3. Restart the agent by running the following command on your VM instance:

     sudo service stackdriver-agent restart
    

Windows

  1. If you use an HTTP proxy, run the following command from an administrator command prompt. This sets the HTTP_PROXY and HTTPS_PROXY environment variables so that the agent can send data using outbound HTTPS:

    setx HTTP_PROXY http://proxy-ip:proxy-port /m
    setx HTTPS_PROXY http://proxy-ip:proxy-port /m
    setx no_proxy 169.254.169.254 /m
    

Determining the agent version

To determine the version of the Monitoring agent on your system, run the following commands on your VM instance:

AMAZON LINUX AMI / CENTOS / RHEL

Run the following command on Amazon Linux, Red Hat, or CentOS Linux:

rpm --query --queryformat '%{NAME} %{VERSION} %{RELEASE} %{ARCH}\n' stackdriver-agent

DEBIAN / UBUNTU

Run the following command on Debian or Ubuntu:

dpkg-query --show --showformat '${Package} ${Version} ${Architecture} ${Status}\n' stackdriver-agent

SLES / SUSE

Run the following command on SUSE:

rpm --query --queryformat '%{NAME} %{VERSION} %{RELEASE} %{ARCH}\n' stackdriver-agent

WINDOWS

There is presently no way to determine the version of the Monitoring agent running on Windows.

Restarting the agent

You must restart the Monitoring agent to pick up changes in configuration files. To restart the agent, use the following instructions.

LINUX

Run the following command on your instance:


     sudo service stackdriver-agent restart

Windows

  1. Connect to your instance using RDP or a similar tool and login to Windows.

  2. Open a PowerShell terminal with administrator privileges by right-clicking the PowerShell icon and selecting Run as Administrator.

  3. Run the following PowerShell command:

Restart-Service -Name StackdriverMonitoring

Upgrading the agent

To upgrade the Monitoring agent to the latest release, use the following instructions:

Linux

To upgrade the agent to the latest version, run the following command:

sudo bash add-monitoring-agent-repo.sh --also-install

To upgrade the agent to the latest point release of a specific major version, run the following command:

sudo bash add-monitoring-agent-repo.sh --also-install \
  --version=MAJOR_VERSION.*.*

Windows

To upgrade to the latest agent release, install the newest agent as described in Installing on Windows on this page. The installer prompts you to uninstall the previous version of the agent.

Listing all agent versions

To list the available versions of the agent, run the following command:

AMAZON LINUX AMI / CENTOS / RHEL

List the available versions of the agent:

sudo yum list --showduplicates stackdriver-agent

DEBIAN / UBUNTU

List the available versions of the agent:

sudo apt-cache madison stackdriver-agent

SLES / SUSE

List the available versions of the agent:

sudo zypper search -s stackdriver-agent

WINDOWS

Installing earlier versions of the agent on Windows is not supported.

Uninstalling the agent

To remove the Monitoring agent and its configuration files, use the following instructions.

After you uninstall the agent, the Google Cloud console might take up to one hour to report this change.

Linux

Run the following command:

sudo bash add-monitoring-agent-repo.sh --uninstall

Optionally, to remove the repository in addition to uninstalling the agent, append --remove-repo to the previous command.

Windows

In the Windows Control Panel, choose Uninstall a program. You should see the Monitoring agent in the list of programs that you can uninstall. You can also run uninstall.exe from the directory where you installed the Monitoring agent.

Information about Google Cloud projects for AWS EC2 VM instances

When the documentation refers to the Google Cloud project associated with your VM instance, for EC2 VM instances, this phrase refers to the AWS Connector project linked to your AWS account.

When you connect your AWS account to a Google Cloud, you create an AWS Connector project. For information about this process, see Collect metrics from AWS accounts.

To access the AWS Connector project for an AWS account, do one of the following:

  • Use the Google Cloud console project selector to identify the projects that match your AWS Connector project naming conventions, and then select the specific project for your AWS account.

  • Identify the Google Cloud project whose metrics scope includes your AWS account metrics and select that project in the Google Cloud console project selector. For this Google Cloud project, go to the Monitoring page and then select the Settings page. The Settings page lists the AWS Connector projects. You can use the Google Cloud console project selector to access the AWS Connector project.

VMs without remote package access

Installing the Monitoring agent requires access to remote package repositories, for both the agent package and (on Linux) its dependencies.

If you are using VPC-SC or a private network, the network configuration might also affect your ability to install agent dependencies from upstream repositories. The agent packages themselves are accessible by using Private Google Access. This can be configured following Enable Private Google Access.

If your VM host's security policy denies access to remote package repositories, we recommend creating a custom VM image with the agent pre-installed and disabling package management in that image.