Troubleshooting

Please refer to the topics below. If still having issues, please open a support ticket.

There are no metrics

• Ensure that composerInstances.yml is correctly specified. Please refer to the Configuration page.

• Ensure that the host running Composer Monitor can access each target specified in composerInstances.yml

For example, assume a composerInstances.yml that looks like this:

- targets: [
  "http://host.docker.internal:8015",   # Gather metrics from THIS host.
  "http://20.20.50.1:8015",             # Gather metrics from 20.20.50.1.
  "http://20.20.50.2:8015",             # Gather metrics from 20.20.50.2.
]

Then from the host running Composer monitor make sure it can access each of these target's Prometheus endpoint, set in Composer. In this example, navigate your browser to:

http://20.20.50.1:8015/metrics

Each endpoint should return Prometheus data:

# HELP composer_frames_processed_total Total number of processed video frames.
# TYPE composer_frames_processed_total gauge
composer_frames_processed_total{label="my-label",version="1.8.8822",hostname="my-hostname"} 32400
# HELP composer_congestions_last_second 
# TYPE composer_congestions_last_second gauge
composer_congestions_last_second{label="my-label",version="1.8.8822",hostname="my-hostname"} 0
...

If not - the specified target's Prometheus port is not available. Please check firewall and other network-related settings.

Firewall settings

Note that Composer Monitoring stack is using a static subnet

networks:
  default:
    driver: bridge
    ipam:
      config:
        - subnet: 172.28.0.0/16
          gateway: 172.28.0.1

If using a firewall, make sure to allow inbound traffic to the Docker container's bridge, such as for UFW:

sudo ufw allow from 172.28.0.0/16

There are no log files

Composer (Desktop and Runtime) must be configured to send log files to Composer Monitor.

• Double check the settings for the failing Composer instance to ensure it specifies its Loki endpoint correctly.
• Double check that the host that is running Composer Monitor exposes the port 3100, so that the Loki ingester is allowed to ingest.

Something is weird after updating

Composer Monitor must be restarted after updating, for the new settings / changes to be applied.

Depending on how Docker was setup on the host, use either command below that works:

docker compose restart

or

docker-compose restart

I have made changes in dashboard X but they are reverted

Composer Monitor will read and use the dashboards as defined by the json files in the Dashboards folder. Any changes are only stored locally while Composer Monitor is running. Restarting will re-construct the Dashboards from the json files.

To persist the changes - refer to the Composer monitor updates article.

I have found a bug or have an idea

We are always happy to receive feedback or hear ideas. Contact RealSprint via our support system.