Make it easy for users to share (query) telemetry data
- Dominant language
- Java
- Stars
- 4.4k
- Forks
- 616
- Avg merge
- 14h 49m
- Merged PRs (30d)
- 206
Description
### Problem Statement
Over the years we've had several discussions about adding more telemetry data to get a better view into what kind of features are used vs. not used.
So far the outcome was that the community tends to view telemetry badly - justified by the privacy impact. Without further insights it can also be challenging to interpret the data.
But there are also data points indicating that users would like to help by sharing their data, especially when it comes to improving performance:
- https://community.cratedb.com/t/very-slow-query-performance/1321/16
Approaches like valve does with the hardware survey also seem to be quite accepted.
Opening this up here to get some feedback on what people think would be acceptable, and what kind of data they'd be willing to share.
### Possible Solutions
To address privacy concerns, whatever we add should probably be an active operation with confirmation, rather than some sort of passive system.
Could also be part of crash or the admin-ui or an extension to https://cratedb-toolkit.readthedocs.io/info/index.html instead of in CrateDB itself, but adding something in CrateDB might have advantages. E.g.:.
Something like a `COLLECT AND SUBMIT METRICS with (duration = '30')` statement that turns on JFR and monitors system metrics, and then submits them.
Or an option for EXPLAIN to make it more query specific.
Contributor guide
Assessment
This issue has not been assessed yet.