Jonathan Shook
Jonathan Shook
We can use the perfteam's grafana instance for the one or more of the grafana dashboards configured into the docker-compose.
In order to support huggingface vector dataset testing for recall and performance, KNN data needs to be prepared for our target datasets. This means using the neighborhood watch project to...
The JMX adapter was used to enable better testing automation with scripting. It needs to be re-implemented on the current adapters API with the features needed for polling, etc.
Name driver spaces the same as the activity alias by default, so that different concurrent activities will have their APIs and contexts automatically partitioned in the runtime. This only applies...
**What is the ideal scenario for using this feature?** A user specifies a local file, classpath resource, or a URL, and they want to be able to specific which of...
During NB bootstrapping, named scenarios are mapped out for the valid named scenario command line forms. When there is an error in this processing, specifically during yaml parsing, the errors...
Allow for the construction of a default workload based on an accessible or provided schema. Initially, this will be provided for the CQL driver and possibly adapted to other drivers...
This initiative is focused on improving the CQL Workload generator. - [ ] #647 - [ ] #648 - [ ] #649 - [ ] #650 - [ ] #651...
We need a focused effort on preparing a simple and repeatable user experience for testing vector stores. Specifically, this should include: - [ ] #2003 - [ ] #2004 -...
A new workflow for ANN vector search testing will use datasets downloaded directly by users for hugging face. Tests should be added for release builds which exercise this in a...