When the Cloud Monitor agent on a host enters the Stopped state, Cloud Monitor stops monitoring that host. Identify the cause and restart the agent.
Issue
The Cloud Monitor agent sends a heartbeat to the Cloud Monitor server every 3 minutes. If no heartbeat is received for 15 minutes, the agent is considered stopped. Possible causes:
The agent cannot communicate with the Cloud Monitor server.
For more information about how to check the network, see Configure network settings for Cloud Monitor.
The agent process exited.
For more information about how to handle an auto-exit of the Cloud Monitor agent process, see What do I do if the Cloud Monitor agent exits automatically?
If the agent still does not send heartbeats, try the following solutions.
Reinstall or upgrade the Cloud Monitor agent
If the agent ran normally before it stopped, reinstall and restart it. If you use the Go or Java agent, upgrade to the latest Cloud Monitor agent for C++:
Back up the Cloud Monitor agent log directory before you reinstall or upgrade the Cloud Monitor agent. The installation process may overwrite the original log files. If the reinstallation fails again, no logs are left for you to diagnose the cause.
Linux: Run the following command to back up the log directory.
cp -r /usr/local/cloudmonitor/local_data/logs /tmp/cloudmonitor_logs_backupWindows: Copy the C:\Program Files\Alibaba\cloudmonitor\local_data\logs directory to another location.
(Recommended) Automatically install the Cloud Monitor agent for C++
Install the Cloud Monitor agent for C++ on a host by using a script
The Cloud Monitor agents for Go and Java are no longer maintained.
View Cloud Monitor agent status and logs
You can only manage the Cloud Monitor agent by using an administrator account (the root user for Linux or the Administrator user for Windows). However, using an administrator account may introduce certain risks, such as system stability or data security issues. Proceed with caution.
Check the agent status and logs to determine the cause of the stop. If the agent process exited due to a bug, submit a ticket and wait for Alibaba Cloud engineers to contact you before troubleshooting.
View agent logs
The console does not provide a query feature for the installation log of the agent. Agent logs are stored only on the local server of the client.
Windows
Log on to the host as an administrator.
Go to the C:\Program Files\Alibaba\cloudmonitor\local_data\logs directory.
Open argusagent.log or argusagentd.log.
argusagentd.log: daemon process logs, including monitoring process startup and shutdown events.
argusagent.log: operational logs of the Cloud Monitor agent.
If the C:\Program Files\Alibaba\cloudmonitor\local_data directory does not exist, the Cloud Monitor agent is not installed on the server. Uninstalling the agent (by running uninstall.bat) only unregisters the argusagent service and does not delete the installation directory. Therefore, a missing directory usually means that the agent was never installed.
Open Services (services.msc) and check whether the argusagent service exists. If the service does not exist, install the agent from the Cloud Monitor console. After the installation is complete, the local_data directory and its logs subdirectory are created automatically, and log files such as argusagent.log are generated in the logs subdirectory. Cloud Monitor does not store agent logs on the server side. All agent logs are stored only on the local server of the client.
Linux
Log on to the host as the root user.
Run the following commands to view agent logs:
cd /usr/local/cloudmonitor/local_data/logs
cat argusagent.log
cat argusagentd.log
argusagentd.log: daemon process logs, including monitoring process startup and shutdown events.
argusagent.log: operational logs of the Cloud Monitor agent.
View agent status
Windows
Log on to the host as an administrator.
Open the Services window.
Press Win+R. In the Run dialog box, enter services.msc and click OK.
Check the status of the argusagent service.
Linux
Log on to the host as the root user.
Run the following command to check agent status:
ps aux | grep argusagent | grep -v grep