Open NielsKorschinsky opened 3 years ago
These spikes are also caused by the pageSize, though the query time of 3 Min is way too high. I've also seen Dashboards with about 17 Minutes. On default load the preferred time is 30 seconds.
Even when looking at query duration per item, vm's is way too high at > 1 second. This may be caused by a missing SPP-Index, but also by the API-Request parameters.
It seems like the
--daily
job has a ~10 min runtime, of which 3 min are soley for the influxdb request to thevms
table. If this is correct and not an mathematic error when generating the statistics, the indices of the vms table should be improved, as well as the kind of data saved and the queried request. There are 6.6k vm's backuped.