Details
-
Task
-
Resolution: Unresolved
-
Major
-
None
-
3.0.5
-
None
-
1
Description
the issue came up in https://issues.couchbase.com/browse/CBSE-8976. Basically, using python and rest directly, the elapsed_time in the metrics is always about 10% shorter than when using the sdk. Of course, as the size of the payload increases, this may be a limiting factor and a hinderance to customers using analytics for big data sort of things.
My guess is we are pulling chunks off the buffer a bit slower than we could. Maybe our bindings? Maybe LCB? Lets a) figure out what happens when the result sets are larger than all the beer-sample, and b) assuming it increases and is worth looking at, isolate where issue is c) fix it or make a ticket