For AI agents: The documentation index is at https://docs.digitalocean.com/llms.txt. Markdown versions of pages use the same URL with index.html.md in place of the HTML page (for example, append index.html.md to the directory path instead of opening the HTML document).

DigitalOcean Monitoring uses the open-source do-agent metrics agent to collect system-level data such as CPU, memory, disk, and load averages. The agent isn’t installed by default. You can either enable it during Droplet creation or install it manually.

If you create a GPU Droplet, selecting Improved Metrics and monitoring (Free) also enables GPU Observability, which adds GPU metrics such as utilization, temperature, power, and throttling.

Install Metrics Agent on Existing Droplet

You can install the metrics agent either by using the automated installation script or by configuring the DigitalOcean package repository directly. The script is faster and simpler, while the repository method gives you more control over package source configuration.

To install the metrics agent on an existing Droplet, use the automated installation script:

1. Go to the DigitalOcean Control Panel, in the left menu, click COMPUTE, and then click Droplets. Then, find and select the Droplet you want to install a metrics agent for, and then in the top-right, click Web Console to open a terminal session.

2. Run the following command to install or upgrade the metrics agent:

   curl -sSL https://repos.insights.digitalocean.com/install.sh | sudo bash

   If the installation is successful, the output shows that the system is updated, required dependencies such as GnuPG and curl are upgraded, and the do-agent service is installed or upgraded without restarting containers, users, or VMs.

3. After installation, run the following command to confirm the metrics agent is active:

   systemctl status do-agent

   If the metrics agent is running successfully, the output looks like this:

   ● do-agent.service - The DigitalOcean Monitoring Agent
       Loaded: loaded (/etc/systemd/system/do-agent.service; enabled; preset: enabled)
       Active: active (running) since Thu 2025-04-17 17:07:44 UTC; 47s ago
   ...

   This confirms that the metrics agent is installed, runs in the background, and starts automatically on boot.

4. After installing and verifying the metrics agent, in your Droplet’s page, click the Insights tab to see your metrics graphs.

After installation, we recommend that you set up resource alerts to help detect issues early and keep your infrastructure healthy.

Set Up Resource Alerts

Resource alerts help you monitor Droplet health and be notified when usage crosses a threshold you define. You can create alerts for CPU, load average, memory, disk, and bandwidth metrics to detect sustained pressure, unexpected spikes, or other changes that may affect performance.

To set up your resource alerts:

1. Go to the Control Panel, in the left menu, click INSIGHTS, click Monitoring, and then in the top-right click Create Resource Alert to open the Create Resource Alert page.

2. In the Select metric & set threshold section, click the Metric Type dropdown list you want to monitor:

   You can select either 1, 5, or 15 Minute Load Average, Memory Utilization Percent, Disk Utilization Percent, CPU Utilization Percent, Disk Read I/O, Disk Write I/O, Public Inbound Bandwidth, Private Inbound Bandwidth, Public Outbound Bandwidth, or Private Outbound Bandwidth.

   For definitions of each metric, see our Droplet Metrics.

3. Click the Rule dropdown list to either choose is above or is below to define the alert condition.

   Use is above when high usage indicates a problem, such as CPU usage exceeding a safe limit. Use is below when low values may signal an issue, such as a drop in bandwidth.

4. Type in the Threshold field to set the threshold value that triggers the alert.

   For CPU, memory, or disk usage, start with a threshold of 70%, since sustained usage above that level may affect performance.

   For load averages, choose a value near or slightly above your Droplet’s vCPU count. For bandwidth and disk I/O, choose a value based on your Droplet’s typical usage.

5. Click the Duration dropdown list to set the duration for how long the condition must persist before the alert triggers. You can choose 5 min, 10 min, 30 min, or 1 hour.

   Use shorter durations for critical services where quick response matters, and longer durations to avoid alerts from brief spikes.

6. In the Select Droplets or Tags section, choose All Droplets to apply the alert to all existing Droplets, or type in the Droplets or Tag field to select a specific Droplet or tag. Only Droplets with the metrics agent installed are available to select.

7. In the Select alert notification method section, choose whether to receive alerts by email, Slack, or both. You can update the notification method at any time.

8. In the Finalize section, either type a name of your resource alert or use the auto-generated name.

9. Click Create Resource Alert.

After you create your new resource alert, it appears in the Monitoring page under the Resource Alerts page.

Uninstall Metrics Agent

Uninstall the metrics agent if you no longer want your Droplet to send metrics data to the Control Panel. After removal, the Droplet stops reporting resource metrics and no longer appears in Insights or alert policies until the agent is installed again.

To uninstall the metrics agent:

Go to the Control Panel. In the left menu, click COMPUTE, then Droplets. Select the Droplet you want to uninstall a metrics agent for, and then in the top-right, click Web Console to open a terminal session.

2. Run the following command to uninstall the metrics agent:

   sudo apt-get purge do-agent

   A prompt confirms that the system found the do-agent package, is ready to remove it. To continue, type Y, and then press ENTER.

   The following packages will be REMOVED:
   do-agent*
   ...
   After this operation, 13.1 MB disk space will be freed.
   Do you want to continue? [Y/n]

3. After confirming uninstallation, the systemd and cron files is deleted, and the metrics agent configuration is removed.

   Note

   You may see warning messages like this:

   Failed to stop do-agent.service: Unit do-agent.service not loaded.
   Failed to disable unit: Unit do-agent.service does not exist

   This means the agent was already stopped and its service file was already removed. The uninstallation still completes successfully.

4. To confirm that the metrics agent is fully removed, run the following command:

   systemctl status do-agent

   If the uninstallation succeeded, the output includes this message:

   Unit do-agent.service could not be found.

   This confirms that the service is no longer installed. If the agent still appears, repeat the uninstallation steps or content support.