2 Replies Latest reply on May 12, 2016 6:58 AM by I LIKE EGGS

    Monitor Polling Gaps via Alert


      Dear All,


      Can anyone help me or guide me how to monitor polling gaps via alert in NPM 11.5.3. Thanks!

        • Re: Monitor Polling Gaps via Alert

          Not sure what you mean here, would you mind elaborating?

          • Re: Monitor Polling Gaps via Alert
            I LIKE EGGS

            Hi rajakamran,


            ideally you dont want polling gaps in your charts or graphs, so if you do see them i would advice to look at some of the following:


            There are a few reason why you could be seeing gaps in Graphs if that is where you are seeing them..


            • Soalrwinds use ICMP and some circumstances SNMP is used for Fault Monitoring, there are times when these Protocols are classed as low priority so these packets get dropped to allow other traffic such as voice..  i have spoken to our network guy and this quality of Service (QOS) is set up by you Network team so check if this is has been setup.


            • your antivirus may lock files used by Solarwinds Job Engine v2 during scanning for malicious files this may cause the Job Engine v2 service to stop and restart causing delays in polling and gaps in data for a poll cycle, Solarwinds does recommend to exclude the programdata\Solarwinds file on the C:\\ (yours might be different and depends on environment)


            • I noticed you said that you have 11.5.3 and this is a new upgrade so you might want to check that the polling settings have reverted back to default which is how frequent you are polling elements/interfaces, Volumes, Applications and Nodes each of these settings are in settings > Thresholds & polling > Polling Settings. A rull of thrum the more granular the detail the more your database will grow
            • Also you cant performance monitor when the node/element is 'down' so you will see gaps in charts at the point the node/element is down


            • you might want to also check the disk queue length (DQL) this is possibly the most important measurement for the SQL server becase it tells you that somthing is waiting to write to the server this could be related to fault or performance monitoring that you might miss due to DQL and if the queue grows you will see these request getting dropped this will also play a role on the Enterprise Operations Console (ECO) your user interface in other words.


            • lol final one if you are using a graph to drill down on a device you might see gaps because the granularity exceeds what is available


            hope this helps