DX Infrastructure Manager

Expand all | Collapse all

Data engine monitoring

  • 1.  Data engine monitoring

    Posted 09-12-2017 09:44 AM

    HI,

       How to monitor my dataengine probe, as i see sometime the dataengine probe stops and queue size is too big. this causes my monitoring system to a halt



  • 2.  Re: Data engine monitoring

    Posted 09-12-2017 10:07 AM

    Hi,

     

    You can explore the CA UIM Hub Queue Statistics Probe

    http://marketplace.ca.com/shop/uim/ca-uim-hub-queue-statistics-probe.html



  • 3.  Re: Data engine monitoring

    Posted 09-12-2017 10:10 AM

    usually this is happening because of a problem with the database.

    You can monitor the data_engine log for failure and timeout messages to alarm on.

    you could also setup the sqlserver probe and do a custom check point to do a row count on the cm_computer_system table or s_qos_data table. if they come back less than 1 send alarm.



  • 4.  Re: Data engine monitoring

    Posted 09-12-2017 10:48 AM

    Indeed, using the Queue Statistics Probe is helpful.

    Monitoring the queue size of the data_engine queue on the NMS hub is the very first thing to do.

    You can use callbacks on the data_engine to get some statistics (get_statistics) and also use some of the following metrics:

    * CPU used by the data_engine process with the processes probe

    * Memory used by the data_engine process with the processes probe

    * State of the data_engine process with the processes probe

    * Size of the queue directory for data_engine containing *.sds files with the dirscan probe

    * Age of the oldest *.sds file in that queue directory with the dirscan probe

    * Number of *.sds files in the queue directory with the dirscan probe

     

    That gives you a good overview.



  • 5.  Re: Data engine monitoring

    Posted 09-12-2017 10:49 AM

    You can also use a small LIA script from: QueueCheck LUA script v2.2 ; this tool will create the needed alarms and qos entries to be alarmed and to do the follow up (with sample list view included)



  • 6.  Re: Data engine monitoring

    Posted 09-12-2017 11:14 AM

    I would say that LUA script run by NAS is less efficient than a dedicated probe for that, but that's another method! J