We poll the following REST endpoint for tunnel statistics for hundreds of thousands of tunnels.
The poll is performed for the last 30m on the hour and 1/2 hour marks.
https://VMANAGE_HOST:8443/dataservice/data/device/statistics/approutestatsstatistics?count=10000&startDate=2021-09-30T22:30:00&endDate=2021-09-30T23:00:00
The probe time is 2m for these tunnels which results in 15 samples for each tunnel. This x15 per tunnel produces a large amount of data. At this time we are fast approaching 30 minutes to complete acquiring all the pages.
This is a problem because the next poll window will have past and polls will start falling behind.
We are only interested in a small subset of fields from each record, so we are flooding the network with data which is ultimately dropped by the client.
At some point Cisco Support recommended the following API.
https://developer.cisco.com/docs/sdwan/#!query-format/simple-query
"...In a simple query, you specify the name or names of the fields whose values you want to retrieve,
the time period over which to retrieve the value, and the order in which to sort the retrieved data...."
We tried this, but Cisco determined there was a larger performance penalty to be paid for using this and advised against using it.
We were also not able to [consistently] acquire the data we were interested in between "startDate" and "endDate" in the original endpoint. This was some time back, so there may have been fixes/enhancements since then to make this solution more viable.
Has anyone had success using "Aggregated Queries"?
Regards,
Fred
