| ---++Metric Collection |
| |
| Metric Collection currently allows to collect the following metrics at process level: |
| |
| 1. Processing time the process spent in the running state in seconds (workflow_end_time - workflow_start_time) |
| 1.Wait time that the process spent in the waiting/ready state. (workflow_start_time - workflow_nominal_time) |
| 1.Number of instances that are failed for a process. |
| |
| To send data to *Graphite* |
| |
| Falcon need to intialize metricNotificationService in startup.properties: |
| |
| <verbatim> |
| *.application.services= org.apache.falcon.metrics.MetricNotificationService, |
| </verbatim> |
| |
| Add following properties for graphiteNotificationPlugin : |
| |
| *Graphite properties* |
| <verbatim> |
| * *.falcon.graphite.hostname=localhost |
| * *.falcon.graphite.port=2003 |
| * *.falcon.graphite.frequency=1 |
| * *.falcon.graphite.prefix=falcon |
| </verbatim> |
| The falcon.graphite.frequency is in seconds and all the time that is being sent to graphite is in seconds. |
| |
| |
| To send data to *Falcon DB* |
| |
| Falcon needs to *!ProcessInstanceInfo* table in the database have a look at [[FalconDatabase]] to know how to create it. |
| |
| Add the following properties in the startup.properties: |
| |
| <verbatim> |
| #*.monitoring.plugins=org.apache.falcon.plugin.ProcessExecutionStatsPlugin |
| </verbatim> |