We are currently discussing on the mailing list whether to upgrade to the 2.0 branch of the e-divisive implementation we depend on: https://github.com/mongodb/signal-processing-algorithms
I create this ticket to track the question, discussion is still ongoing. Findings so far:
- Adds two algorithms: 1) compare 2 distributions for similarity or distance, 2) outlier detection. 1) exists independently in Otava, and 2) it's unclear why or what is the use case?
- Multi-variate support in e-divisive: The algorithm supports consuming multiple metrics together, which should improve accuracy. While this would be great to have, I should not my memory of reading about this is that improvements were marginal. (But real.)
- I seems someone from MongoDB perf team might join the mailing list. It will be interesting to see what they say!
We are currently discussing on the mailing list whether to upgrade to the 2.0 branch of the e-divisive implementation we depend on: https://github.com/mongodb/signal-processing-algorithms
I create this ticket to track the question, discussion is still ongoing. Findings so far: