We're troubleshooting an issue with slow SNMP responsetimes for a certain type of equipment.
In the protocol under General Parameters / Communication info section there are som values that may be of interest, but haven't found an explanation how it's measured and the time base (i.e. real time, or average over an interval?)
How is device RTT measured (is it response-time per GET? or something else?)
What is device iterations? Does a device message drop equal a snmp get request with no response within set timeout value or no response after x number of retries?
When you click the diagram symbol you get a graph that I don't understand really...
Hi Roger,
The parameters have a brief description of what they are if you double-click a cell in the column you want to know more about.
In short, most of these values are 5-second averages using windows of 1 second. In other words, every second a window is closed, and its metrics are averaged out with the 4 preceding windows.
These values are calculated using the data provided by the Bit rate calculations feature, for which the linked help page provides more details. The Device Iterations in particular is the amount of SNMP requests that have received an SNMP response within the aforementioned window.
The diagram symbol in the header of a table is a generic feature that allows you to view a heatmap and/or histogram of all values in that column. In the case of the Communication Info table, what you are seeing are likely histograms, which show the occurrence of each value in the column.
Hi Ben, yes, the Round Trip Time is the average response time. In practice, we accumulate the total SNMP RTT (value at index 10) and accumulate the iterations (index 7) within a time window. Then when updating the table, we divide the total SNMP RTT by the amount of iterations to get the average RTT.
Maybe also the Stream Viewer can give you some additional insights in what is happening? Connecting to an element using Stream Viewer | DataMiner Docs
On a lower level, a lot of people use Wireshark to inspect traffic. Wireshark · Go Deep
I’ve been using both the stream viewer logging in combination with wireshark and seen that we have snmp requests without response and also sometimes very slow response times. Problem is that sometimes it’s quite ok and sometimes it’s worse. I figured that if I had a better understanding of what the values mean I could trend on them and get a long-time picture of how bad it is.
Edit: Specifically, it’s our Tredess FS DTTV transmitters which seems quite slow and sometimes miss a request, but a couple of them are really bad for unknown reason and we may have to see if we can tweak the timeout settings for those as a workaround.
We also have a PING driver to track the basic performance of an IP infrastructure. Maybe worth to set one of those up and track the PING performance with those devices? Or do you have other devices in the same environment that don’t have any of those problems, and which more or less rule out IP network issues?
And the RTT Floris. This is the average time for DataMiner to receive a response when it does an SNMP Get?