Closed

Closed due to inactivity. Received 17 votes with last vote on 02 Sep 2019.

Enable a toggle within the gui to display interface utilization above 100% - integrate a success center fix

https://support.solarwinds.com/SuccessCenter/s/article/Polled-traffic-over-100-not-showing-in-graphs   - this article exists to display interface utilization when it's above 100%. You can indeed set custom bandwidth to display within your 100%, but if you have a burstable interface with a set bandwidth this idea doesn't work. The problem is, you have to change this on every single poller.

I'd this to be a checkbox in polling settings so that you don't need a support article to be able to modify this.

The article says (and what I'd like to be a GUI checkbox/toggle):

Resolution 2 (Steps must be completed against all polling engines):

  • Warning:  Save a copy of the original SolarWinds.Interfaces.Collector.dll.config file to your local drive as a backup file, in case you need to roll back later.
  • SolarWinds strongly recommends that you only edit the SolarWinds.Interfaces.Collector.dll.config file as instructed. Any additional modifications may result in system performance issues or may create an error state.
  • Log in to your Orion server.
  • Go to C:\Program Files(x86)\SolarWinds\Orion\Interfaces\
  • Open SolarWinds.Interfaces.Collector.dll.config in a text editor.
  • Replace the following line:
    <add key="ValidateCounters" value="True" />
    with
    <add key="ValidateCounters" value="False" />
  • Restart SolarWinds services. Refer to Start and stop Orion services.
  • No worries good sir! emoticons_happy.png I would wonder about use cases in both scenarios for this.

  • Your analysis has merit.  I can't disagree.

    Until such time as it's available, I don't know an alternate other than going the manual route.  But I'll be watching here with interest for other ideas!

    Somehow I was completely confused and thought this was a Discussion, not a Feature Request.  My apologies for that, designerfx.

  • This is a FR rschroeder​.

    pastedImage_0.png

    The problem is, you're having to set custom interface bandwidths that may not be true. Here's an example:

    You have a gigabit interface, your actual bandwidth is 20mb, but you can burst up to (infinity, who knows). A custom bandwidth doesn't work for that. You want your alerts at 20mb but you want your bandwidth to clearly show when you are >100%, as that will show up on the homepage under interfaces with high utilization. Does it show >100%? Is it 100%? Yes to both.

    I get changing this threshold, but I wouldn't want to deal with that across >1000 WAN devices, especially if those devices are having custom bandwidth programmed into them which is supposed to be their non-burstable speed. Otherwise I'm doing some unfun admin work emoticons_sad.png

    Without this toggle, when it goes above 100% you get zero. However, the toggle requires editing a file and stop/start services, so a checkbox and/or moving this into something for the database would be more logical.

  • It's an interesting problem to have.  Where one has more than 100% of the rated bandwidth hitting an interface one might expect errors and discards, depending on how QoS or ToS is configured.

    Each time we set up a new WAN we monitor all active ports on the firewall or L3 switch or router doing the connecting at both ends.  We also deploy QoS policing to limit the traffic flow to 100% of the contracted rate.  If we don't do this the WAN provider will, and the traffic will be dropped instead of queued.

    At the same time we also configure the NPM monitor for the WAN interface to specify the contracted throughput rate.

    This combination of things serves to prevent the problem you seem to be reporting, where more than 100% of the available bandwidth is being consumed.

    I'll grant you that it would be useful to see if and when any interface is sending/receiving more than 100% of its rated throughput.

    One could set up an alert/threshold to fire each time more than 100% is seen, and use it as an action item to investigate the problem and correct it.  This would also work for burstable network connections--you'd immediately know if/when more than the low-end rate is exceeded.  Your ISP or WSP is undoubtedly already doing this so they can bill you each time that burstable link exceeds the base threshold.  They will be able to show you this information--although "be willing to show you" could be the fly in the ointment.  But if they're charging you more for exceeding the base level, they must be able to show you the proof for the billing.

    I'd consider changing the NPM interface monitoring level to show the true highest-amount of bandwidth that could be generated on a burstable link.  Then set the alert threshold for that interface at whatever percent is your base contracted threshold.

    For example:

    Let's say you have a WAN circuit that should do the job at 20 Mb/s most of the time, but might temporarily burst to 200 Mb/s during a failover / traffic-re-routing event.  You're paying $100/month as long as the throughput doesn't exceed 20 Mb/s.  But as soon as it hits 21 Mb/s or higher you start paying a higher rate.

    In this screen shot you can see how this alerting is set up.

    pastedImage_0.png

    1. I selected Custom Bandwidth

    2. I entered in the maximum allowed bandwidth

    3. I used the Override Orion General Thresholds option to;

         a. Put the interface alert into Warning when more than 10% (or 20 Mb/s) is used on that link.

         b. Put the interface alert to Critical when throughput hits 80% (or 160 Mb/s) is used on that link.

    Then simply ensure your Alerts fire off something that will catch your attention.  An e-mail, something on the front page of NPM, a text to your Service Desk, etc.

    Is this an option that might work for you until your vote gets enough attention to be implemented?

    Somehow I was completely confused and thought this was a Discussion, not a Feature Request.  My apologies for that, designerfx​.