DX Application Performance Management

Expand all | Collapse all

Large ts_defect_meta_values table

Jump to Best Answer
  • 1.  Large ts_defect_meta_values table

    Posted 05-18-2017 07:32 AM

    Hi all,
    I have a problem with APM 10.2 on Oracle Db; I'm using CEM and the Db is now about 800 Gb and it seems that there are too many old data; for example in ts_defect_meta_values is 300 Gb and I see defect from 17/01/2017 and ts_defects is 7 Gb

     

    I have closed all the incident and I have set :

     

    under Setup -> Incident Settings

     

    Delete Defect after 1 days

     

    Under Setup -> Domain

     

    Make Users Inactive after: 35 days
    Keep Hourly User Statistics for: 1 days
    Keep Daily User Statistics for: 14 days
    Keep Weekly User Statistics for: 27 weeks
    Keep Hourly User Group Statistics for: 1 days
    Keep Daily User Group Statistics for: 14 days
    Keep Weekly User Group Statistics for: 27 weeks

     

    Any advise ?

     

     

    Thanks,
    Daniele



  • 2.  Re: Large ts_defect_meta_values table

    Broadcom Employee
    Posted 05-18-2017 07:48 AM

    Hi Daniele

    Addition to the steps you have mentioned  you need to check few points

    1:  Why there is too much defect is creating ? and do you really need them

         please check your threshold value  and the condition you have defined  to create a defect

    2:  I am not sure when you have enabled those settings but it will be effective after  midnight

        if tomorrow also no improvement then

    3: You need to enable the MOM  +Tim collector service log in debug mode  and check if there is any

    issue with defect aggregation or not 

     

    Hope this will help you ..

     

    Thanks



  • 3.  Re: Large ts_defect_meta_values table
    Best Answer

    Broadcom Employee
    Posted 05-18-2017 08:22 AM
      |   view attached

    Dear Daniele:

            I am attaching the Database maintenance Tech Note because you may have otehr db issues as well.

          Per Olympia's note there are two sides to this issue

     

    THE INPUT 

        - Monitoring too many or unneeded transactions

        - Low defect thresholds

        - Capture comprehensive defect details enabled 

        - Request and Response body size larger than default

     

    THE OUTPUT

        - Many incidents and defects

        - Growing database 

       - Slow CEM GUI

       - Future headaches

     

    Keep track also if 35 days is too low and you are getting too many inactive users

    Thanks

    Hal German

    Attachment(s)