Is there a way I can monitor a "heartbeat" for the Yellowfin application?
There are a few ways one could monitor to verify they Yellowfin application is still up and running.
1. The PersistentWebserviceSessionMonitorTask - this task runs every five minutes. It produces a log entry in the yellowfin.log file:
INFO (SystemTaskManager:?) - Success instantiating class: PersistentWebserviceSessionMonitorTask
which could be used to verify the heartbeat. If the time exceeds 5 minutes, it may indicate an issue with the application.
2. This same task can be monitored in the 'TaskSchedule' table of the Configuration Database. This will show the last run status as well as last run time, stored in the database. NOTE: This will not work for individual nodes of a clustered environment, as the nodes share a Configuration Database.
3. Monitor the connections to the Configuration Database. If you are able to monitor active connections to the Yellowfin Configuration Database, you can effectively determine whether the application is up. This is a good option for clusters, as you will have open connections from multiple sources.
4. My favorite option utilizes web services to poll the application. The attached jsp can be hosted on the application server to verify whether the Yellowfin application is up. Hitting this jsp page in your browser issues a "TEST" Web Service call, which simply returns "SUCCESS". If this is returned, it generates this output in the browser. This option can be configured on each application server (if clustered) and will have a very low resource overhead for polling.
Comments have been locked on this page!