Docs

Torch Log Service

Data collection FAQ

How to troubleshoot log collection exceptions (container)

After using LogCollector to collect container logs, if logs are not successfully collected, refer to the steps in this document to troubleshoot.

1. Check whether the host group heartbeat is normal

View the host group heartbeat status in the TLS console. For details, see View heartbeat status..

If the number of normal heartbeats in the target host group does not equal the number of nodes, refer to How to troubleshoot host group heartbeat exceptions (container) to troubleshoot.

If the target host group has no heartbeat, reinstall LogCollector. For details, see Install LogCollector (DaemonSet deployment).

If the heartbeat of the target host group is normal and the quantity is correct, proceed to the next step for further troubleshooting.

2. Check whether the collection configuration is correct

Confirm in the TLS console that the LogCollector collection configuration has been created and is correct.

If the collection configuration has not been created, create the collection configuration first. For specific operations, see Collect container logs.

If the collection configuration has already been created, check the following:

The file path in the collection configuration is the absolute path inside the container.

The symbolic link path inside the container is not used as the collection path.

The container filtering rules in the collection configuration meet your collection requirements.

Notice

The Label in the general container collection rules refers to the container-level Label, not the Label in a Kubernetes Pod. You can temporarily remove the container filter rule. If normal logs can be collected at this time, it indicates that the collection Exception is caused by incorrect container filter rule configuration.

The log file matches the collection configuration correctly.

By default, a log file can only match one collection configuration. If logs in the file need to be collected multiple times, refer to Collect multiple log files or standard output simultaneously.

If there are no issues, proceed to the next step for further troubleshooting.

3. Check whether the collection configuration is bound to the host group

In the host group details page, check whether the collection configuration has been bound to the host group. For details, see View host group.

If the host group is not bound, refer to Change the host group for the collection configuration to complete the configuration.

If the host group is already bound, proceed to the next step for further checking.

4. Check the container to be collected

Confirm that the container to be collected meets the following conditions:

Ensure that the container survives for at least 10 seconds so that LogCollector can collect logs.

LogCollector supports two container runtimes: Docker and Containerd. Other container runtimes are not currently supported.

When using Docker as the container runtime and collecting container standard output, ensure that daemon.json in the Docker configuration file includes "log-driver": "json-file".

5 View error information

Check the LogCollector logs for any error information. The LogCollector client records key information as well as all WARNING and ERROR logs. You can view the logs in the LogCollector installation directory. The log file paths are as follows:

LogCollector version	Installation scenario	File path
LogCollector v2.2.0 and later versions	Host scenario	/usr/local/logcollector/logs
LogCollector v2.2.0 and later versions		Container scenario	/usr/local/container/logcollector/logs
Logcollector versions before v2.2.0	Host scenario	/usr/local/filebeat-7.12.0/logs
Logcollector versions before v2.2.0		Container scenario	/usr/local/container/filebeat-7.12.0/logs

Notice

To view LogCollector version, see 2. View software version.

After completing the above checks, if the issue is still not resolved, submit a ticket and provide the key information found during troubleshooting in the ticket.

Last updated: 2026.06.25 16:39:46