dimitry12
commented
8 years ago @powerc9000 my idea was to post that to CloudWatch for now.
The development of the idea, together with @bluetechy was to:
- accumulate enough data-points,
- then use anomaly-detection algorithms.
Anomaly detection, in the spirit of Machine Learning, would recognize if current Travis-node got slower (or some other factor affecting all tests in a given run) or it's the specific test that ran anomalously slow.
Many games have a leader-board or high score board. It would be cool to create a dashboard displaying the commits that scored better than others using the results from Travis-CI which could then be saved over to an S3 bucket for analysis.
I know @dimitry12 had some interesting ideas on how to create something like this.