Component StartupΒΆ

Note

All project executables are installed with a version suffix not mentioned here in the documentation. E.g. rsb-host-monitor actually needs to be called as rsb-host-monitor2.0.

To start all components of the monitoring pipeline, follow these steps (assuming that RSB has been configured properly):

  1. Start the timeseries database backend:

    InfluxDB

    $prefix/opt/influxdb/influxdb -config=/path/to/influxdb-config.toml
    

    Graphite

    1. Start the carbon daemon:

      $GRAPHITE_ROOT/bin/carbon-cache.py --nodaemon
      
    2. Start the webapp:

      $GRAPHITE_ROOT/bin/run-graphite-devel-server.py --port=${PORT}
      

      Decide on a port you want to use.

  2. Start one rsb-host-monitor instance per computer. Pass the intended measurement cycle time to the --cycle option. A typical call might be:

    rsb-host-monitor --cycle 5000
    
  3. Start rsb-process-monitor instances. Depending on how you automate your application startup, you can opt to start one instance per process or use a single instance fir multiple processes on a host. The processes to monitor are specified via positional arguments or the --pid option. A typical call might be:

    rsb-process-monitor --cycle 5000  -s -r -i -f -u -n a_name $PID
    
  4. Start a single rsbperfmon-db-adapter instance. A typical call will be:

    rsbperfmon-db-adapter -a -b {graphite|influxdb}
    

    Use the appropriate backend for the -b option. Additional flags for the connection to the backend might be required.

  5. Start the Grafana Dashboard server:

    $GRAFANA_PREFIX/bin/grafana-server -config=your-config.ini -homepath=$GRAFANA_PREFIX
    

    This assumes that GRAFANA_PREFIX points to the directory containing the Grafana installation.

  6. Open the Grafana dashboard in your browser and select one of the previously imported dashboards configuration (see Configuration of the Grafana Dashboards).