DX Application Performance Management

  • 1.  Observed % sign in agent node

    Posted Mar 22, 2017 03:53 AM
      |   view attached

    Hi All,

     

    We observed there are one extra node connected with %1.What is meaning for this(%1) and if i want to remove this node what should i do to remove from investigate ?

     If this is important node then how i can modify and remove only %1 sign ?

    This is for weblogic  application and we are using CA APM 10.1.I have attached problem area in this question.Can you please help me on this. 

     

    Regards

    Vinay Nayak 



  • 2.  Re: Observed % sign in agent node
    Best Answer

    Broadcom Employee
    Posted Mar 22, 2017 09:14 AM

    Hi Vinay:

         This is due to your Agent having Autonaming being enabled whether Java or .Net . See Java Agent Naming - CA Application Performance Management - 10.5 - CA Technologies Documentation and https://www.ca.com/us/services-support/ca-support/ca-support-online/knowledge-base-articles.tec1546308.html 

     

    for details.  Please let us know if it helpful. You may ask additional questions as well.

     

    Thanks

    Hal German



  • 3.  Re: Observed % sign in agent node

    Posted Mar 23, 2017 02:38 AM

    Hi Hal,

     

    How i can stop %1 from agent level.

     

    Regards

    Vinay Nayak



  • 4.  Re: Observed % sign in agent node

    Broadcom Employee
    Posted Mar 23, 2017 06:59 AM

    Dear Vinay:

    This was covered in teh document. You mentioned your agent is Weblogic which autonaming is on by default. I am going to include all agents since someone else may refer to this in the future

     

    To specify an agent name using Java system property, on the Java command line, supply the desired name using this property:

    -Dcom.wily.introscope.agent.agentName=

     

    Do the opposite here for JBOSS

    https://docops.ca.com/ca-apm/10-5/en/implementing-agents/java-agent/install-the-java-agent/configure-application-server-to-use-the-java-agent/jboss#JBoss-ConfigureJBoss7forAutonaming

     

    Agent Naming Properties
    You can configure properties to obtain the Java agent name for application servers.


    https://docops.ca.com/ca-apm/10-5/en/implementing-agents/java-agent/java-agent-properties


    introscope.agent.agentAutoNamingEnabled
    This property specifies whether agent autonaming is used to obtain the Java agent name for supported application servers.

    Values: true or false

    Default: Varies by application server.

    Example: introscope.agent.agentAutoNamingEnabled=false

    Notes:

    Agent autonaming is enabled by default for WebLogic and WebSphere.


    https://docops.ca.com/ca-apm/10-5/en/implementing-agents/net-agent/properties-reference-net-agent

     

    introscope.agent.agentAutoNamingEnabled
    Specifies whether agent autonaming will be used to obtain the .NET Agent name for supported application servers.

    Values: True or False

    Default: True

    Example

    introscope.agent.agentAutoNamingEnabled=true

     

    That should cover it. Please mark/comment if helpful.

     

    Thanks

    Hal German



  • 5.  Re: Observed % sign in agent node

    Posted Mar 23, 2017 07:24 AM

    Hi Hal,

     

    We are using weblogic server and and Autonaming is false.

    introscope.agent.agentAutoNamingEnabled=false .I guess it should not reflect ? is there any other way to remove it ?

     

    Regards

    Vinay Nayak



  • 6.  Re: Observed % sign in agent node

    Broadcom Employee
    Posted Mar 23, 2017 07:37 AM

    Dear Vinay:

    Your logs (I believe don't need DEBUG/VERBOSE) will give you a clue as to why 

     

    There is a very good Knowledge Document on other things to check

     

    https://www.ca.com/us/services-support/ca-support/ca-support-online/knowledge-base-articles.tec604522.html 

     

    including "Try extending the delay for Agent Auto Naming in IntroscopeAgent.profile"

     

    failing that, please open a Support Case

    Thanks

    Hal German



  • 7.  Re: Observed % sign in agent node

    Broadcom Employee
    Posted Mar 23, 2017 09:02 AM

    Hi Vinay,

     

    Did this happen after you installed it or has this been running fine and then you just noticed this extra instance in the GUI?  Double check the metric tree and see if both instances are identical.

     

    If you are seeing one extra node there are a few reasons for this.

     

    1. You may have another agent instance using the same profile setup running on the same server.

    2. Your application may be spawning off another process which is being picked up by our agent.

    3. Your agent may have been restarted or restarting too quickly before the EM realizes it has disconnected.

     

    Are you using windows or linux?

    If windows, open up task manager.  Make sure that you add Command line and PID to the list of columns to view in here.  Sort by name and see where java.exe is coming from.  Double check

     

    If linux, perform the same using ps -ef | grep java

     

    This may show two different paths or same path but different process name, i.e. your app could be named Medrec but it may be spawning off another process named ContactDomain or something like that.

     

    If however you only see one weblogic process per above, then you would need to check the agent logs to see if there were any disconnects and reconnects in a short period of time.

     

    You could also try stopping weblogic, wait until both instances go grey in the GUI and then start it up again.  If both show up, then its most likely another process being spawned off in the same java process.  From here you would investigate your startup script for weblogic, generally named startWebLogic.cmd or startWebLogic.sh depending on your OS.

     

    Thanks,
    Matt



  • 8.  Re: Observed % sign in agent node

    Broadcom Employee
    Posted Mar 28, 2017 11:16 AM

    Hi Vinay,

    Hal and Matt have already covered the possible causes and resolutions, I believe extending the delay for Agent Auto Naming in IntroscopeAgent.profile should help.

    You don't need to check/compare all the metrics, just check the Agent\ProcessID metric and correlate them with the running processes in the system.

    If you still have any doubt please open a support case, include agent logs + appserver logs + startup script + result of processes (including the pids)

    Thanks,

    Sergio