4 Replies Latest reply on Aug 25, 2017 9:41 AM by ecklerwr1

    Bandwidth Utilization Alerting

    spyfly

      I am curios to how others are alerting on bandwidth utilizations.

       

      We currently are alerting when the usage is over 95% for 10 minutes.

       

      Any thoughts on input would be great, we are trying to make sure we are in the normal range of this type of alerting

       

      Also while we are at it let's talk about what is the recommendation or standard on increasing bandwidth to a site based on a 95 percentile as well.

       

      Thanks

        • Re: Bandwidth Utilization Alerting
          ecklerwr1

          A lot depends on your situation.  You've already caught onto one of the most important aspects of this... the time it must exist before alerting!  I have some circuits that totally peg at night while things like NAS backups or large data transfers occur.  If you're doing QOS this can help but you need QOS working end to end and to be very well organized or maybe QOS isn't right for you.

           

          Reporting is another area where watching your bandwidth utilization month after month can really be helpful with sizing your circuits... that's what I do.

          • Re: Bandwidth Utilization Alerting
            d09h

            Do you monitor flow information?  If a Cisco network, NetFlow would be immensely helpful to know how much different types of traffic contribute to the aggregate (backups/ streaming video/ mission related traffic). Depending on what you see, you may decide to QoS bandwidth hogs, or ensure that backups happen at an agreed up acceptable time.

            • Re: Bandwidth Utilization Alerting
              spyfly

              We do have QOS, not 100% sure it has been updated over the years.

               

              Our main issue is during business areas as we service the public in many locations and when a circuit is pegged it slows down that service, so the idea is to try to be proactive to a point when we see a link peg at 100% we look at scrutinizer (netflow tool) and figure out the issue and reach out to the end users to have them limit what they are doing, example pulling data from remote cameras, youtube, etc..