Closed caseycesari closed 8 years ago
Does the network graph distinguish between transfer time and waiting time? Do we still have a sleep
in the stub endpoint? Would be interesting to see that breakdown, but ultimately we may want to revisit #303. 6 seconds is a long time for no real processing.
Interesting. I went back to get more information, and now repeating the steps above is pretty fast. I was testing mmw-dev
and tried a number of different watersheds, all of which were slow. But using the 1 sq kilometer tool was fast.
Maybe it was something related to the environment getting setup or a briefly slow connection?
Closing for now.
Sure - depending on when you were testing it, our internet at the office was really slow. During the Stroud call, we had to phone in because the VoIP stuff wouldn't work. Chip said that we were at ~25% capacity at the time, could have been related.
Each endpoint takes a few seconds even with the fake endpoints.
I think this is because we are posting large BLOBs of GeoJSON.
To reproduce, select a the Lower Delaware Subbasin, and note the request times on the analyze and modeling pages.