The benchmark is composed of a serie of a dozen of Gatling simulations. Those simulations cover important use cases of a content platform such as mass import, create, read, search, browsing, update and delete of content.

They are part of the Nuxeo source.

The simulations are driven by a Jenkins job.


Each simulation plays a scenario with a concurrency that is part of the report, there is no pause (i.e. think time) between requests, (except for the simulation Benchmarks mixing actions).

Mass import

This simulation use the nuxeo-platform-importer to generate random document with an attachement. The attached binary file is stored in S3 and the fulltext is extracted and indexed. The import is run on a single Nuxeo node.

Create document using REST

Create documents using the REST API. The documents are imported from a CSV file, there is no file attachement.

Read using REST

Get a random folders and documents using the REST API with various metadata.

Search using REST

Performs fulltext search using terms from a CSV file.

Using the web UI, view a random folder and a document in it, view all document tabs.

CRU+Delete using REST

CRUD is Create/Read/Update/Delete of document, this is done on top of previously imported documents.

Benchmarks mixing actions

Mixing different simulations: Navigation JSF, Read and Update using REST. This simulation uses pauses between requests.

Reindexing repository

Time to drop and recreate the Elasticsearch index, documents are loaded from the repository.