skip to navigation
skip to content

esbench 0.2.0

Elasticsearch performance benchmark tool

Latest Version: 0.2.3

Tool for benchmarking performance of Elasticsearch nodes.

The two primary uses are for capacity planning (guessing how much oomph you need to do what what you think you need to do), and for performance tuning (trying out various index, mapping, and query settings in a consistent and reproducible manner).

An Elasticsearch index is composed of a set of 1 or more Lucene indexes (designated as primary and replica ‘shards’ by ES). A single Lucene index is the basic unit on which indexing and search operations are executed, and so the performance of individual Lucene indexes largely determines the performance of a cluster.

The basic approach is to create an index with 1 primary and no replica shards (a single Lucene index), load it with data, and periodically run representative use patterns against it, recording observations, adding more data until the performance drops below acceptable levels.

This tool comes with ‘batteries included’ (ie large sample data set, downloaded on demand). See the file, or even better, the project’s github page.

File Type Py Version Uploaded on Size
esbench-0.2.0.tar.gz (md5) Source 2014-01-16 35KB