I have two separate indices for same data. I would like to compare the querying speed for the two indices using a python script. I am using the 'took' variable in my response as a metric.
The thing I want to avoid is the use of cache by the query. I have set the parameter 'request_cache' as False but I guess that does not seem to be working. Is there something wrong I am doing?
It doesn't make any sense to measure the speed of something with the cache disabled. This doesn't tell you anything useful, because in real life the cache is enabled and is a critical factor affecting how the system performs.
Rally is a good way to benchmark your cluster and helps you avoid some common pitfalls. There is also a recorded presentation about common mistakes when benchmarking that would be good to watch.
Thanks for the response.
Actually I wanted to measure the improvement of the new index that I created with a well defined mapping. So all I wanted to do was see if that new mapping helps me .(The new mapping has a lot few fields that are indexed and the original mapping had all the fields as indexed.)