Icon

Comments have been closed on this page. Please use AppMon & UEM Plugins forum for questions about this plugin.

Overview

 
The APMaaS Metrics Grabber plugin utilized the web services available from Compuware's APMaaS offering to return summarized test information from a specified APMaaS Backbone test script. At this time, only backbone tests are supported.

The plugin uses the APMaaS WSDL Interface to query the following metrics:

  1. Availability
  2. Response Time
  3. Connection Time
  4. SSL Time
  5. DNS Time
  6. First Byte Time

     

Plugin Details

Plug-In File

com.compuware.APMaaS.scriptmetrics_1.4.5.jar

Author

Michael Beemer (michael.beemer@dynatrace.com)

Joe Hoffman v1.3.0 (joseph.hoffman@dynatrace.com)

dynaTrace Versions

>= 5.6

License

dynaTrace BSD

Support

Not Supported

Known Issues

 

Release History

2013-4-24 Initial Release

2013-5-2 v1.3.0 Release.  Fixed runtime bugs, improved error messaging. Repackaged acccording to plugin guidelines and naming conventions

2015-2-20 Fixed a number of bugs and added a mechanize to handle multiple requests at the same time

Icon

In order to utilize this plugin, you must have at least read access to the APMaaS Network and the specified active backbone script.

Configuration

  • Enter the full name of the backbone script
  • Enter your APMaaS login account information
  • Enable debugging for a more verbose log
Icon

The polling interval should be the same or longer than the APMaaS script interval

Results

The following screenshot shows a dashboard displaying some of the measures queried by the monitor:

Things to consider

  1. Only works on backbone tests
  2. Gomez only allows a limited number of sessions to be opened simultaneously. (Handled better in V 1.4.5)
  3. Numbers probably won't match exactly to what you see on Gomez Networks because of how the data is being summarized.
  4. The polling collector must have access to the internet

Troubleshooting

  • Enable the debug flag and check the collector logs
  • Look at the root of the dT collector folder for "{your_script}.xml" and make sure the data is accurate
    • MID
    • Time stamps
    • Script name
  • Delete the XML and rerun the test
  • Add a comment to the thread if you run into any other issues.

What we need from you!

  1. Testers
  2. Ideas on how to make it better
  3. Logs from any failed test (Please enable debugging)

Contribution

Feel free to contribute any changes on Github

  1. Anonymous (login to see details)

    PLM- What would be needed to make this happen? I know that OOTB in BSM, backbone test are the only supported. Is there something preventing from pulling data from PLM and Last Mile Tests?

     

    1. Anonymous (login to see details)

      Hi Mike,

      I had trouble getting a valid response from the Gomez web services when I tried to pull back PLM data.  It was something I was going to work on if there was enough interest.

      Thanks,

      Mike

  2. Anonymous (login to see details)

    Anyone been working on this?  We see an opportunity to leverage this for a situation with PNC.

    1. Anonymous (login to see details)

      Hi David,

      I can help you would with any issues you have with the plugin.  Please respond to this thread or email me directly with any questions.

      Thanks,

      Mike

  3. Anonymous (login to see details)

    Any issues with this an dT 5.5?  I'm getting this following error...

    2014-01-05 17:35:08 WARNING [JavaUtils@New Gomez Script Metrics_0] Unable to find required classes (javax.activation.DataHandler and javax.mail.internet.MimeMultipart). Attachment support is disabled.

    2014-01-05 17:35:09 INFO [dynatrace@New Gomez Script Metrics_0] Failed because of a JAXB exception!

    2014-01-05 18:03:22 INFO [dynatrace@EasyTravel TXN - 20-Sep-2013 18:16:21 - 2 - FF Agent_0] Failed because of a JAXB exception!

     

     

  4. Anonymous (login to see details)

    Is this still valid for Dynatrace 6?

    We would like to get Backbone single URL test metrics.

     

    Thanks

    Grant

  5. Anonymous (login to see details)

    Just to add a +1 to the interest around including PLM stats. 

    1. Anonymous (login to see details)

      Hi Sean,

      I know there is a ton of room for improvement.  Honestly, I would like to completely redesign the whole plugin at some point.  The Gomez team is putting the finishing touches on a REST interface that should make it significantly easily to get all the information you're looking for.  It's my goal to have a fully updated plugin shortly after that's released.  Unfortunately, I can't give any ETAs at this time though.

      Thanks,

      Mike

      1. Anonymous (login to see details)

        I'm very interested in this as well.  We would like to create a Dashboard in dynaTrace that essentially shows Green or Red based on the Alerts we send out from Gomez.  For instance, some of our tests send emails if 3 out of 4 nodes fail.  So if this happens, we'd like to turn the dashboard entry for that test Red.  Other tests, it has to fail on 100% of the nodes.  So it varies.  Ultimately, I think we'd like the StopLight chart and have it Green for no current failures, Yellow if there's a failure but not to the threshold of alerting, and then Red if it hits our Alert threshold.

        We'd also like to see this be a dynamic measure.  Setting up a measure for every test within Gomez would become very tedious.  So if the plugin could pull all the tests and create a dynamic measure, that would be idea.  I could also see value in having a dashboard that shows the tests that are Active verus Inactive.  So when a test hits a maintenance window and the test is not active, this is shown on the dashboard.


         

  6. Anonymous (login to see details)

    Hi Michael,

    No problem, I wasn't looking for an ETA, I just wanted to voice my interest. 
    Nice work on a great plugin.

    Thanks

     

    Sean

  7. Anonymous (login to see details)

    Hi Michael

    I imported this into a 6.1 server and got the same error as David above.  Was there any update to this?

    Any ideas as to the cause?

    2015-02-18 13:09:33 INFO [APMaaSMetricsMonitor@Test APMaaS Script Metrics_0] Found Script name: http://www.macquarie.com/au - 18-Feb-2015 0:26:42 - 4 - FF Agent
    2015-02-18 13:09:33 INFO [APMaaSMetricsMonitor@Test APMaaS Script Metrics_0] an error has occured
    2015-02-18 13:09:33 INFO [APMaaSMetricsMonitor@Test APMaaS Script Metrics_0] Script created correctly
    2015-02-18 13:09:33 INFO [APMaaSMetricsMonitor@Test APMaaS Script Metrics_0] Failed because of a JAXB exception!

     

    Thanks

    Grant

  8. Anonymous (login to see details)

    Hi Grant Engelbrecht,

    Look for an update tomorrow.  I just fixed a couple of issues and added some additional logging.  I'll post it as soon as I finish testing.

    Thanks,

    Mike

  9. Anonymous (login to see details)

    Thanks Michael

    All good for v6.1 with your latest fix.

    Will there be any enhancements for retrieving metric for more 1 BB script at a time?

     

    Thanks

    Grant

  10. Anonymous (login to see details)

    Hi Grant Engelbrecht,

    I don't have any plans to extend a single monitor to poll data about multiple BB scripts at this time.  What would the use case be?  You can add as many monitors as you want and each one will poll it's own BB script.  The latest update should handle the limitation of only being able to have a single session token out at a time.  If it notices that a session is already open, it will pause for a random amount of time and try again.  It will continue to do this for a couple of attempts before giving up.

    Hope this helps.

    Mike

  11. Anonymous (login to see details)

    Mike,  any chance you can provide more detail on what the Hosts configuration should look like?

  12. Anonymous (login to see details)

    David Jones, the host configuration is more or less a hack.  It does nothing in the background but is a required parameter for plugins in dynatrace.  Simply fill it in with any host you want.

  13. Anonymous (login to see details)

    We're running a 6.1 POC and getting the following error logged:

    2015-06-19 16:46:59 INFO [APMaaSMetricsMonitor@New APMaaS Script Metrics_0] Checking the collector for Siebel Prod OpenUI Backbone - FF Agent.xml
    2015-06-19 16:47:21 SEVERE [APMaaSMetricsMonitor@New APMaaS Script Metrics_0] RemoteException: ; nested exception is:
        java.net.ConnectException: Connection timed out: connect
    2015-06-19 16:47:21 INFO [APMaaSMetricsMonitor@New APMaaS Script Metrics_0] an error has occured

    This plugin has proved its usefulness during another recent POC. Thanks.

     

  14. Anonymous (login to see details)

    Greg Speckhart,

    Does the collector you're using have access to the internet?  Try to run the plugin from a different collector if possible.

    Please let me know if you're still having issues and I'll take a look.

    Thanks,

    Mike

  15. Anonymous (login to see details)

    I was told by the customer the Collector has access to the internet – I will verify and follow-up with you. Thx.

  16. Anonymous (login to see details)

    Greg, Make sure it is using the correct collector.  I recently ran into an issue where we installed the plugin and were getting results like you described.  Turned out their were multiple collectors and we had not selected the correct one.

  17. Anonymous (login to see details)

    Script name with spaces  - No data polled

    While playing with this plugin, i found that if the gomez script name contains space, there will be no data showing in dynatrace, if i change the script name to not contain space, then data comes in. Can anyone confirm this?

    1. Anonymous (login to see details)

      Hey Xiaoping, 

      Is this still occurring for you?  I believe this is due to the Collector being Linux and the plugin therefore creating a "script name.xml" file on the collector host to track and parse the test results.  Since Linux obviously doesn't like spaces, I believe it is presenting issues within the plugin.

      Michael Beemer - Can you please confirm this is the case?  Also, any chance on getting a solution implemented so that the plugin will work on Linux collectors?  Unfortunately, despite the web dashboard and new integration, AppMon still doesn't contain custom measures for Synthetic tests availability.

      Thanks!

      Josh

  18. Anonymous (login to see details)

    Hello,

    I am having issues with this plugin. I have confirmed that Synthetic Monitoring is properly enabled in the dynaTrace Client, and the APMaaS Script Metrics plugin is active and configured, but I am still not seeing the APMaaS Script Metrics options under System Monitoring when trying to create a measure for charting. Is there anything else I need to configure to get this working?

    Thanks,

    Annmarie

  19. Anonymous (login to see details)

    Hi Annmarie,

    Have you configured the monitor under your system profile?  This plugin works independently of the Synthetic Monitoring integration within dynaTrace.  You should be able to test your configuration to make sure it's able to pull back data.

    Thanks,

    Mike

    1. Anonymous (login to see details)

      That was my missing piece, thanks!

  20. Anonymous (login to see details)

    Hi Michael, hope you are well.

    I've got this plugin up and working now (had to make some tweaks to make the call through a proxy), however I just want to confirm my thinking on a couple of things around dashboarding.....(I know you said metrics will never tie up exactly, but I want to make sure I'm as close as possible)

    Presume I have a Gomez  BB/Ops dashboard setup, and my time filter is set to one hour. The dashboard alerts if the availability drops below 90% (allowing for a single failure in a one hour rolling window before triggering an alert).

     

    What is the best way to create a similar dashboard using your plugin?

    Chart availability as a meter / alert chart, but set the timeframe and data resolution to one hour? - I seem to get odd results when doing this. 

    E.G. if i have resolution set to 1hour, i will see 100% availability, with the same time range if i change the resolution to 15minutes i can see 87.5%, change to 1minute and i see 85.71%


    Looking in Gomez the actual availability over the last hour is 91.67 and only one test has failed. 

    Should I actually be setting the polling interval to one hour - thus getting a full hour of data to base an availability calculation from? - Doing this makes it to slow to alert on in production though. 


    Thanks in advance.


    Sean. 

     

     

    1. Anonymous (login to see details)

      Hi Sean,

      It's best to run the monitor at the same interval as your tests.  I say that the data may be different than what you see in the portal simply based on timing.  For example, the monitor could be pulling data in right before a few test locations report their test status.

      I would assume based on your comments above that the reason you're seeing seemingly invalid data is because of the charts default aggregation.  Double check that you're reporting the average and not last.

      Thanks,

      Mike

    2. Anonymous (login to see details)

      Hi Sean,

       

      Any chance you could share your updated version with proxy support?

       

      Thanks!

      KR,
      Kristof 

      1. Anonymous (login to see details)

        Hi Kristof,

        Happy to share but my only word of warning is that my version is now heavily customised for my use case (the data is not even collected / aggregated in the same way as the original plugin).

        I think Mike has recently created a new version of the plugin which is probably worth trying out before you look at mine. - From looking at his commits, it seems he has also added proxy support (smile)

        https://github.com/beeme1mr/Dynatrace-Synthetic

         

        If you do want to check mine, it is here: https://github.com/smc87/Dynatrace-APMaaS-Metrics-Grabber-Plugin

        But like i say.... very specfic to my use case at the minute (we use it for realtime alerting of issues with gomez scripts and then forward the event via snmp to our alert presentation system).

  21. Anonymous (login to see details)

    Hi Mike,

    Thanks for that, I'll have a play around with the polling interval. I am definitely charting average.
    I wonder if the issue is that more than one test has been pulled back in some of the polling intervals. 

    Cheers,

    Seán

    1. Anonymous (login to see details)

      Shoot me an email if you're still having trouble.  It should be a fairly easy fix.

  22. Anonymous (login to see details)

    Hi Mike,

    Fantastic plugin, thank you for developing this! What are the chances of you adding response time by step to the measure results? (smile)

    P.S. Let me know if you want me to be a beta tester for any early releases you're working on for this!

  23. Anonymous (login to see details)

    +1 for retrieval of PLM stats! (smile)

    And of course, great work on the plugin Mike!

  24. Anonymous (login to see details)

    It would be great to have Synthetic Alerts added to the Grabber - I know its been mentioned before but just thought I'd add another comment to nudge the effort.

  25. Anonymous (login to see details)

    Does the grabber support last mile synthetic transactions?  that is the only type of ST which we run at Raymond James.

  26. Anonymous (login to see details)

    Hey Mike,

    Thanks for updating the plugin to support last mile tests!  I'm currently in the process of implementing this at multiple sites. 

    Hopefully a quick question, does the plugin support pulling data from Benchmark Tests?  I'm currently getting "Unexpected Error" within the new plugin and "Script Not Found" errors within the old APMaaS plugin when trying to load data from Benchmark scripts.  Based on these messages, it appears that the script may need to be owned by the account in order for metrics to be retrieved, is that correct.

    Just let me know, thanks again!

    Josh

    1. Anonymous (login to see details)

      Yes, you need to own the test.

  27. Anonymous (login to see details)

    Hello,

    I am using Dynatrace 6.2.5 with the APMSaaS Script plugin 1.4.5.  We have 5 scripts running 1 minute apart from each other.  They are frequently and randomly failing.  I have set the debug logging to finer and I see the following error in the log:

     

    2016-02-29 13:48:00 FINE [ContextFactory@P DWIP Payment Center - FF Agent_1] Property com.sun.xml.internal.bind.XmlAccessorFactoryis not active. Using JAXB's implementation
    2016-02-29 13:49:00 FINE [ContextFactory@P DWIP Sign In - FF Agent_1] Previous message was repeated 5 times.

    I tried deleting the corresponding XML files on the collector, they were recreated on the next run, but the issue is still occurring.  The XML looks like the following.  Just making sure this is how it should look:

     

    <?xml version="1.0" encoding="UTF-8" standalone="true"?>
    -<config><scriptName>P DWIP Billing and Payments - FF Agent</scriptName><scriptId>22361245</scriptId><transactionType>TRANSACTION</transactionType><runtime>2016-02-29 19:27:05</runtime></config>

     

    Does anyone have any idea why this would occur?  Thank you very much for your help.

     

    Thanks,

    Tom 

    1. Anonymous (login to see details)

      Tom Davison, could you send me an email regarding this issue?  My email is michael.beemer@ruxit.com.