Learn about the Wavefront Apache Spark Integration.

Apache Spark Integration

Apache Spark is an open-source cluster-computing framework. This integration explains how to configure the Wavefront proxy to listen for Graphite data, and how to configure Spark metrics for Graphite. In addition to setting up the metrics flow, this integration also sets up a dashboard.


Apache Spark Setup

Step 1. Configure Wavefront Proxy to Listen for Graphite Data

  1. If you have not already done so, install a Wavefront proxy.
  2. On the server running your Wavefront proxy, open the file /etc/wavefront/wavefront-proxy/wavefront.conf. Uncomment and configure the following properties as required:
  3. Save and close wavefront.conf
  4. Restart the proxy: sudo service wavefront-proxy restart.
  5. Tail the proxy log file to make sure there are no errors: tail -f /var/log/wavefront/wavefront.log.

Step 2. Configure Spark metrics

  1. Create or edit the metrics.properties file in the conf directory of your Spark installation.
  2. Enter the following snippet:
     #Enable Graphite
     *.sink.graphite.host=[proxy address]
     # Enable jvm source for instance master, worker, driver and executor
  3. Use these two properties in the snippet to specify your Wavefront proxy:
    • *.sink.graphite.host=[proxy address]
    • *.sink.graphite.port=2003
  4. Set up your Spark node name:
    • *.sink.graphite.prefix=spark.[hostname]
    • Note: Replace . in the hostname with _. For example, enter spark.prod_host1 instead of spark.prod.host1.
  5. Distribute the metrics.properties file to all your Spark nodes.
  6. Restart your master and slaves nodes.