DX Unified Infrastructure Management

  • 1.  Getting a large amount of QOS data for various sources

    Posted Aug 27, 2009 07:04 PM
    We want to be able to get a type of QOS_DATA for all servers defined by
    either Origin or User_tag2 and either query it directly or export it to
    a file.

    So bsaically,

    If we have Mailserver defined for user_tag2, we would want to find
    a way to get the processor queue length for all servers with user_tag2
    set to mailserver.

    I noticed in the Wizards there is one to export QOS data, but all
    I saw was either select one by one or sometimes it lets you do all.

    How would we go about doing that?
    Can I do this with the nsa?


  • 2.  Getting a large amount of QOS data for various sources

    Posted Aug 27, 2009 08:05 PM
    The current database model does not allow you to utilize the user_tags as items for a query.  We are working on a new NiS model that will open up for this.  However, you can do this with qos and origin.

    Here's a sample script for the NAS (it has NiS integration built into it).  You could ofcourse also use the NSA but you would require a connection string.  The script will list out the matching origins and print the number of data rows for the last week.

    Carstein

    ---8<-------8<-------8<-------8<-------8<-------8<-------8<----

    origin="change-me"

    -- Open the NimBUS Information Store (NiS)
    database.open("provider=nis;driver=none;database=none")
    rs = database.query("SELECT * FROM S_QOS_DATA WHERE qos='QOS_PROC_QUEUE_LEN' and origin='"..origin.."'")
    if rs ~= nil and #rs>0 then
       for i,r in ipairs(rs) do
          printf("source: %s, target: %s, table_id: %d, table: %s",r.source,r.target,r.table_id,r.r_table)
          data = database.query ("SELECT * FROM "..r.r_table.." WHERE table_id="..r.table_id.." AND sampletime > DATEADD(week,-1,GETDATE())")
          printf("    rows of data: %d",#data)
          -- write the data to a file or something
       end
    end
    database.close()



  • 3.  Getting a large amount of QOS data for various sources

    Posted Aug 27, 2009 08:38 PM
    excellent, thanks! this should help me get started


  • 4.  Getting a large amount of QOS data for various sources

    Posted Oct 20, 2009 01:48 AM
    Has anyone does this via the nsa?


  • 5.  Getting a large amount of QOS data for various sources

    Posted Oct 20, 2009 01:52 AM
    sorry i should clarify
    Is there a way to use the API/NSA without doing a direct database call?


  • 6.  Getting a large amount of QOS data for various sources

    Posted Oct 20, 2009 01:30 PM
    Andrew,

    I'm still not 100% clear on what you want to achieve.  All qos data is posted as messages on the NimBUS, thus allowing for subscribers to collect the metrics as they come in.  I posted a solution in the 'Customer Upload Archive' forum that used NSA to collect QoS messages off the bus and keeping the last sample in a in-memory hashtable so I could do top-n reporting/summary on the QoS data.  But you could very well store the QoS data that you need in another database.

    On the other hand, if you want to extract large chunks of QoS data from a historical perspective (minutes,hours,days,months) then your source of information is in the NiS.  The NSA also supports the database class, but without the "provider=nis" shortcut.  You'll need to add a connection-string (encrypted or not) as part of your database.open().

    Carstein


  • 7.  Getting a large amount of QOS data for various sources

    Posted Oct 20, 2009 08:40 PM
    basically I was asking if there was a way through the API/NSA, to get a large amount of QOS metrics WITHOUT doing a database call. Perhaps calling the variable server's query callback, or any other means. We would love to have a way to get QOS data without having to call the databse directly for those developers who do not have database access.


  • 8.  Getting a large amount of QOS data for various sources

    Posted Oct 21, 2009 02:23 AM
    I was able to get it via a Nimrequest to the variable server callback for query.
    Fantastic!