Time series of Netrounds measurement data are stored in a round-robin database, where older data is progressively consolidated into lower resolutions.
Resolution for monitorings
In each of your monitorings, the Test Agents periodically collect measurement data in 10-second intervals, then compile the results into individual measurement reports and send them to the cloud server. The measurement data can be said to have a "resolution" of 10 seconds.
In the database, the 10-second resolution is retained for data from the last 12 hours. Older data is consolidated as indicated in the following table:
To enhance your understanding and ease your interpretation of the various graphs in your monitorings, let us consider an example:
During a 10-second interval, there have been two seconds where the Test Agent has measured a high level of packet loss or delays. The measurement report will then indicate 2 out of 10 seconds with error = 20% errored seconds. In the Netrounds GUI, this is indicated as a red bar representing an ES level of between 10% and 49%.
In the GUI, you can zoom to different time intervals and drill progressively deeper into the measurement history. As you zoom the graphs in and out, you might see the colors (= error levels) in the graph changing. This is because the measurement resolution changes with the time interval displayed. For instance, changing the zoom level from "last 15 minutes" to "last 24 hours" changes the resolution from 10 seconds to 30 seconds. Assuming no other errors than the two errored seconds just mentioned, these no longer correspond to an ES level of 20%, but rather to 6.7%.
Assume now that the loss you see in a graph or table is 0.8% during a 10-second interval, when the resolution is 10 seconds. Assume further that this has triggered an errored second, although your ES threshold is set to 1% loss. How can that be? The reason is that the errors may have occurred in bursts; at one extreme, you might have had 8% loss during a single second, triggering an ES for that second, and no errors in the rest of the 10-second interval.
Resolution for tests
Tests in Netrounds use a higher resolution than monitorings. For tests of duration up to 1 hour, the resolution is 1 second; for tests up to 2 hours in length, the resolution is 2 seconds; and so on for longer tests.