JMX documentation

Has anybody come across a "real" documentation regarding the JMX operations available on the probes/servers?

For example, the "Mbean: KPI:service=Dashboard. Method: viewDiscoveryProcessingStatistics" seems to return a treasure of useful performance information, but the official JMX documentation simply states something like "displays discovery processing statistics" (pretty obvious) but further neglects to explain the meaning of the different data fields (columns).  Although the naming of the data fields gives a hint, it's impossible to match the returned stats to anything reported in the AdminUI.

For example "Number of CIs" could be anything from created/updated/discovered/processed/... CIs.

  • Hello,

    The jmx methods are documented here Indeed not every method may be documented.

    viewDiscoveryProcessingStatistics jmx is identical to jmx exportDiscoveryProcessingStatisticsToExcel. Only the output format is different.
    In both cases the data in log files are read and parsed -> cmdb.reconciliation.audit.log*

    These methods is usually invoked to check discovery throughput and spot bottlenecks.

    Number of CIs represent the total CI number processed in a timeframe(a day usually)-adds+updates+deletes of CIs. The sheet name should be clear enough to understand the context what Number of CIs represents(all CI number in a day, or by job or probe in a day).

    If there are more jmx methods which you believe require more details is recommended to submit an idea exchange.

    Hope it helps.

    Best Regards,
    Gabi Medan

  • Thanks for trying, it provides the smallest snippet of increased awareness but adds to the confusion because of the referal to a log file (another aspect for which there is no documentation).

    My issue with the lack of JMX and Logging documentation comes from the fact that I'm trying to setup monitoring processes to collect insights into what is going on in UCMDB Server and the Probes (because we had issues in production).  There is a lot to be found in JMX and Logs (and Admin UI) but many JMX ops seem to aggregate data in not shown arbitrary timeframes (making it useless because it can't be checked against or correlated to other data returned in similar JMX ops and/or Admin UI).

    MF First Line Support is no help since they have no clue either (which I understand) and it's difficult and time-consuming for all parties involved to get to the level engineers/developers "just to understand" the data.

    I have simple "idea exchange"s :

    - Provide detailed documentation on JMX operations (explanation of functionality, explanation of all required/optional input arguments, explanation of output).

    - All JMX operations should offer choice of output format: HTML and JSON

    - Provide detailed documentation on log file content (explanation of every single field and meaning and possible reasons for error and warning level messages).

    - Implement a standard logging format across all log files