In the graph below note that there is no min/max bps for the tunnel interface (green data points between 9am and 3pm. Anyhow have a sense why SNMP would be grabbing data but not presenting those specific datapoints in that time period?
Do you have the Orion Application template deployed with SAM to watch for jobs lost and jobs queued?These 2 counters are critical to watch for this kind of issue, and for general Orion health.
I would be lost without watching these 2 counters. They are the cornerstone of Orion health.
Take a look at the min/max bps for that interface in isolation, either from the interface details page widget, or using PerfStack. The device appears to be displaying SNMP data from 9am and 3pm for other elements, so it could be that for some reason data was simply not polled, and Orion is extrapolating the plot lines for the missing intervals.
Also, of you were to check the min/max widget on the interface details, and click export, you have an option to view the chart data, which will immediately show if there were any polling gaps, as an alternative to delving into the database tables for it.
Finally, If you look at it in PerfStack, you have other options available for dealing with 'smoothing' when there are gaps, rather than just assuming a straight line, as per the standard widget graphs, which can look better if you need to present Orion graphs to other teams.
