This page lists the various benchmarks we run for different projects within Elastic. They are created by the developers themselves, and focus on the most important performance metrics.
We run benchmarks oriented on spotting performance regressions in metrics such as indexing throughput or garbage collection times.
Indexing 11 million location documents and running various full text queries (match, function_score, …) and aggregations.
Indexing 60 million location documents and running various geo-based queries (polygon, bounding box, distance, …).
Indexing 2 million AOL queries and running percolator queries against them.
Full text oriented benchmark measuring indexing performance and running queries and aggregations on a corpus of scientific papers.
2015 taxi rides in New York measuring the performance for a large highly-structured data set.
Questions and answers from Stackoverflow measuring the performance for a nested data set.
Web server logs from the 1998 football world cup to evaluate the performance for server logs.
Weather data from NOAA measuring the performance of range fields.