Docs

Elastic Compute Service

Logs and monitoring issue

How to resolve the failure in collecting monitoring data by Cloud-monitor-agent of ECS instances?

This topic introduces how to troubleshoot the failure of viewing monitoring information of ECS instances caused by the Cloud Monitor plugin.

Scenarios

When the Cloud Monitor plugin is not installed, is outdated, or malfunctions, monitoring metrics of ECS instances may be missing. You can follow this topic to troubleshoot the issue and restore metric display.

Issue confirmation

Execute the following command on the instance to confirm whether the Cloud Monitor plugin is installed.

systemctl list-units --type=service

If the Cloud Monitor plugin is installed, as shown in the picture below, follow the troubleshooting guide for further operations.

If the cloud-monitor-agent service is not found, complete the plugin installation.

tip

For Windows operating system instances, refer to Install the plugin on Windows instances.

For Linux operating system instances, refer to Install the plugin on Linux instances.

Troubleshooting

Missing operating system monitoring

If the monitoring metrics of your operating system are abnormal or missing, troubleshoot based on the following content.

Execute the following command in the instance to view the running status of the Cloud Monitor plugin.

systemctl status cloud-monitor-agent

If the service status of the Cloud Monitor plugin is active(running) (normal operation), as shown in the picture below, submit a ticket or contact your account manager for technical support.

If the plugin is not in the active(running) (normal operation) status as shown in the picture below, refer to Installing the plugin to update the Cloud Monitor plugin.

Follow these steps to confirm the repair result:

Execute the following command to confirm that the Cloud Monitor plugin is in the active(running) status.

systemctl status cloud-monitor-agent

Execute the following command to verify whether the version of the Cloud Monitor plugin meets expectations.

LD_LIBRARY_PATH=/usr/local/cloud-monitor-agent/lib /usr/local/cloud-monitor-agent/cloud-monitor-agent version --config-file /usr/local/cloud-monitor-agent/config.yaml

If any result of the previous two steps is still abnormal, submit a ticket or contact your account manager for technical support.

Missing TCP connection monitoring

If metrics of TCP connection monitoring are abnormal or missing, troubleshoot as described in this section.

warning

If TCP connection monitoring is enabled, the performance of the instance TCP bandwidth will be impacted to a certain extent.

To avoid data collection failure caused by an in-band monitoring component of an outdated version, we recommend upgrading the plugin as guided by the console prompts. Or, refer to Installing plugins.

Confirm whether the image used by the instance supports viewing TCP connection monitoring metrics. For the list of supported images, see Usage instructions.

If the image supports metric viewing, continue with the following troubleshooting.

Install the Cloud Monitor plugin with reference to Installing plugins.

During plugin installation, the systemctl daemon-reload command is executed automatically. To avoid business interruptions, please do not perform any operations on the instance that may cause conflicts.

Execute the following command to check whether the Cloud Monitor plugin service is running normally.

systemctl status cloud-monitor-agent

Example of normal operation:

Execute the following command to enable the TCP connection service.

sed -i 's/EBPF: true/EBPF: false/' /usr/local/cloud-monitor-agent/config.yaml

Execute the following command to restart the Cloud Monitor plugin service.

systemctl restart cloud-monitor-agent

When all steps are successfully performed, wait a few minutes. To obtain technical support when metrics are still missing, submit a ticket or contact your account manager.

Missing process monitoring metrics

If only metrics of "Network bytes received/sent" and "Network packets received/sent" under process monitoring are missing, confirm:

The keep-alive time of the TCP connection exceeds 30s, and processes have performed network I/O.

The operation flow matches the process monitoring descriptions.

Last updated: 2026.06.18 17:48:25