Join GitHub today
GitHub is home to over 31 million developers working together to host and review code, manage projects, and build software together.
Sign upPromQL: Covariance and correlation functions #2615
Comments
This comment has been minimized.
This comment has been minimized.
|
In my opinion that's getting into heavy statistical analysis, and a dedicated tool such as R would be a more appropriate way to manage that. |
This comment has been minimized.
This comment has been minimized.
adamdrake
commented
Apr 16, 2017
•
|
I've been thinking about this as well. It would probably be helpful to have this functionality for the reasons mentioned. @brian-brazil, can you help me understand, from a project perspective, when something becomes heavy statistical analysis? There is other functionality (e.g., Holt-Winters) present which seems more complicated than correlations, so I'm not sure where the line is drawn. |
This comment has been minimized.
This comment has been minimized.
|
Holt winters currently is the line, and I'm not sure we'd even add something else that complex. |
This comment has been minimized.
This comment has been minimized.
|
R is a great tool, though it means having to use a separate tool (and learn it if I didn't know it). With the addition of the remote read functionality, it's feasible that PromQL could be used increasingly for queries that were perhaps less obvious for the alerting use case. |
mattbostock commentedApr 13, 2017
Are there any plans to support covariance and correlation functions in PromQL?
It could be useful for post-mortem analysis to find correlations that could explain the impact or possible root causes of an incident.