DX Unified Infrastructure Management

 View Only
  • 1.  Development of dyn. dashboards with SDP/UMP - more information?

    Posted Jun 08, 2010 02:32 PM

    Hi,

     

    I'm trying to develop new dynamic dashboards in the SDP/UMP enviroment.

    The topic grouping of robots/servers are clear.

    During the development I had a lot of questions about using of variables, adressing of centralized probes a.s.o.

    Is there any writen information which predefined variables exists within the dynamic dashboards?

    In the standard dyn. dashboard I found variables like:

           $robotip, $robotname, $os_major, $os_minor, $os_description, $source .........

    Are there any more 'hidden' variables?

    What does this means: DD_Override?

    What does this menas: <DataProbe robot="/SDP/van01/van/01" ?

    Are there any other StoredProcedure for the dashboards as "spn_dp_TemplateTabelExpander"?

    How to adress a probe at the nimbus server for a nimRequest? If I adress my custom probe on the nimsoft server I got errors. The dashboard_engine.log shows that the adress of my defined nimRequest was modified by ....(i don't know) - the robotname within the nimadress-string "/domain/hub/robotname/probe" was exchanged to the robotname of the dynamic robotname.

     

    Where is the crack who can tell me (and all other participants) how to develope dynamic dashboards?

     

    After many hours of development I believe that it is a great miracle if you get the expected result ;-)

     

    Regards,

     

    Olaf



  • 2.  Re: Development of dyn. dashboards with SDP/UMP - more information?

    Posted Jun 09, 2010 10:47 PM

    I actually just answered this one recently:

     

    http://forum.nimsoft.com/t5/General-Development/Creating-New-Dynamic-Dashboards-in-UMP-SDP/td-p/2813

     

    Hope that helps!

     

    Dustin



  • 3.  Re: Development of dyn. dashboards with SDP/UMP - more information?

    Posted Jun 10, 2010 11:28 AM

    Hi Dustin,

     

    thanks - I know your doc very well. It helps me in front of my actual problem.

     

    My actual problem is to build some dynamic dashboards/views for different views at the systems (system-view, database-view [ms-sql and oracle]) and using a centralized custom probe (e.g. print a cpu-usage sorted list of processes). I can manipulate the XML-Code of the dashboards (sometimes it is easier to hack the XML code as edit the GUI) but if I don't understand the usage and automatic setting of variables, stored procedures and override rules I can not explot the potential of the dynamic views.... and that would be a shame...

     

    Regards,

     

    Olaf



  • 4.  Re: Development of dyn. dashboards with SDP/UMP - more information?

    Posted Jun 11, 2010 02:57 AM

    I can't get this to work either it's pretty frustrating in that using EC takes about 5 minutes to set something up.  I have the groups showing in SDP but getting a dashboard associated with them has been a chore.  It appears the document is a little outdated as well maybe not meant for sdp 2.6. 

    I am not using User tags either but just custom groups for cisco qos_policies and call manager metrics.



  • 5.  Re: Development of dyn. dashboards with SDP/UMP - more information?

    Posted Jun 15, 2010 07:53 PM

    FYI - I've emailed this to the head developer for the UMP, I know he's traveling right now for work, but hopefully he'll respond once he has a moment.

     

    Dustin



  • 6.  Re: Development of dyn. dashboards with SDP/UMP - more information?

    Posted Jun 16, 2010 01:24 AM
    Thanks Dustin,

    I hope we get more information about the mechanism behind the dashboards. Every Input are welcome.

    Olaf


  • 7.  Re: Development of dyn. dashboards with SDP/UMP - more information?

    Posted Jun 22, 2010 06:18 PM

    Hi everybody!

    You are absolutly right, the creating of dynamic views templates are probably the most difficult aspect of using SDP!

    When starting designing SDP we did not believe it would become such a popular feature (in fact we mainly believed this would be a feature used by Nimsoft staff).

     

    I hope the following comments will help you understand how the templates work and what they can offer. We're in the process of redesigning the support for 'dynamic views' (and will of course also support the existing) and when this is completed the corresponding support in SDP/UMP will be documented in a much better way.

     

    Anyway here are the answers to your questions.

     

    Hidden variables

    Yes, there is a list of ‘hidden’ variables, which variables depend on type of dynamic views dashboard template you are working on.

    The list of variables for a template matches the columns in a table in the NIS database matching the type of dynamic views template as follows, just put a $ in front of the column name to make it a variable:

    • Server templates have variables which match the columns in the table grp_server
    • Connectivity templates variables which match the columns in the table grp_connectivity
    • Interface templates variables which match the columns in the table grp_interface

     

    Alarm filter

    DD_Override is a value you will see for some fields in the alarm filters for dynamic views templates and is mainly there to indicate that the probe dashboard_engine will overwrite whatever is in this field when the template is instantiated to a dashboard. For example the field Robot in a server template will be overwritten with the content of  the cell defined by column robotname and the corresponding row in grp_server.

     

    Probe requests

    <DataProbe robot="/SDP/van01/van/01" is actually rather misleading. When a probe request is defined in a dynamic views template, the address of the probe is just a placeholder. Typically when you design the template you will chose the probe offering the information you would like to display. As long as you can access the probe from DashboardDesigner you can chose any such probe, f.ex. the local cdm probe or another cdm probe running on a different robot/hub. As long as the probe can offer the set of callbacks you require you’re fine since the probe dashboard_engine will, when template is instantiated to dashboard, use the address defined by the server (/domain/hub/robot/) to address the correct probe. The address /SDP/van01/van01 only exists because this was the server the template was originally developed against at Engineering in Nimsoft.

    The explanation above should also answers the question regarding addressing probes, all probe requests are modified to point to the probe related to the node/server the mapped against.

     

    spn_dp_TemplateTabelExpander

    No, there is sadly no other procedures like spn_dp__TemplateTableExpander_01. However the  good news is that you can easily write your own. The purpose of these procedures is to allow the dashboard_engine to create the required sql for the related table widget the instantiated dashboard. The dashboard_engine will at that time edxecute the procedure (using any supplied arguments, and a few which is supplied by dashboard_engine by default). Before starting to create your own procedures you should make sure you understand how spn_dp__TemplateTableExpander_01 works. You can then create say spn_dp__TemplateTableExpander_02 and use this in the widget’s datasource.

     

    Regards,

       Guttorm



  • 8.  Re: Development of dyn. dashboards with SDP/UMP - more information?

    Posted Sep 28, 2010 12:23 AM

    It appears that when we try to use variables (for example $source) in queries to DBs other than the NIS DB, the parameter/variable is not actually parsed / replaced with real data as we would expect it to be. Running a trace on the DB shows that the query executing actually has the "$source" string in it, rather than the value for the source.

     

    Can someone confirm that this is the case, and if there are any plans to fix this?



  • 9.  Re: Development of dyn. dashboards with SDP/UMP - more information?

    Posted Nov 10, 2010 12:18 AM

    Hate to bring up an old topic, but I'm having trouble with the alarm filters on the interface traffic dynamic dashboard.

    The status icon has an alarm filter containing 'Message = DD_Override $interface' but this does not seem to be pulling up the alarms for that interface. What I'm thinking is needed is an alarm filter containing '*'s around the interface name so that it searches for 'Message = /.*DD_Override: $interface.*/'.

    Is this the right approach? Does anyone else have issues with the interface traffic dynamic view not accurately reflecting alarms?