4 Replies Latest reply on Sep 21, 2012 4:06 AM by Ismo

    Netflow polling interval and NTA charts

    WinskiTech711

      Using NTA 3.8, whenever I look at NTA charts, every source/endpoint/etc dips down to 0 at a 30 minute interval. Our charts end up looking like jagged teeth for lack of a better description. As the traffic on my network doesn't completely die every 30 minutes, I'm assuming this is either a polling interval issue or something that can be "smoothed out" by changing the way the chart looks.

      I've googled around and for the life of me cannot find where to change the netflow polling interval, or how to change the interval on the charts so it looks smoother. Can someone help me with this?

      Thanks,

        • Re: Netflow polling interval and NTA charts
          WinskiTech711

          I'm incredibly surprised I've turned up nothing on Google about Netflow polling intervals, I thought Netflow was fairly popular...

          In any case, maybe a different approach is required. Assuming Netflow is being transmitted all the time, is there some way to change the interval that Solarwinds reads that info?

          Submitting a trouble ticket about this issue.

          • Re: Netflow polling interval and NTA charts
            Jesquitin

            By default Cisco and other vendors will exports long lived packets every 30 minutes, which explains what you are eperiencing.  By adding the following command should resolve your issue and level off the peaks and valleys:  ' ip flow-cache timeout active 1'.   This command will ensure all flows are exported every 1 minute.

            To clarify.  Netflow is a push technology and thus not polled.  Depending on the type of flow the will export them at different times. 

             

            For example:  Telnet connections are exported once the session has terminated.

             

            Hope this helps

             

            Thanks,

              • Re: Netflow polling interval and NTA charts
                Ismo

                I have this "cache timeout active 1"-command included (actually our Cat6500 supervisor is new model, it uses new netflow protocol which contains little different style of configurate netflow...) But anyway... command is there, but I still get netflow graphs in 1 hour interval. How can I make interval more frequent in ORION NETFLOW SETTINGS? I have newest version of NTA and NPM.