Telemetry
Help us improve Onyx
Overview
The term telemetry here refers to the collection of anonymous data that is limited in scope and for the purpose of improving the quality of a piece of software.
This page describes our philosophy towards telemetry, what kind of data is collected and how to easily opt-out of all data collection.
Why do we collect telemetry
As a software that uses a combination of deep-learning models and LLMs as well as interfacing with a variety of external SaaS tools for data collection, Onyx’s performance is strongly dependent on a multitude of factors that we cannot predict. Telemetry helps us understand how the system is performing in real life situations and allows us to improve specific bottlenecks within the flow.
We appreciate your help in making Onyx a better software!
What data is collected
| Field | Description |
|---|---|
| Latency | Because latency depends strongly on deployment choices including hardware, etc., this let’s us know if/where the product is too slow for a significant number of users. |
| Connector Failures | This tells us if any data connectors are flaky and need additional retry logic etc. Only the connector type is reported. |
| Random Deployment UUID | We use this to understand general usage and also to differentiate between widespread issues vs repeat issues. |
| Number of active users | No other information about users are collected. This is used to understand the scale that each deployment is expected to handle. |
How to opt-out of telemetry
You can opt-out of usage data collection by setting the DISABLE_TELEMETRY environment variable to True.
For example, with docker compose deployments, simply update your .env to include: