Closed jortizcs closed 11 years ago
@jortizcs Have you done anything on this? I can take care of this if you want
Maybe this shows a misunderstanding on my part, but would it be possible for the anomaly to change over time? If so, we ought to have a way to "freeze" an anomaly as well, if you want to send a URL.
I'm not sure I understand your concerrn... An anomaly is characterized by two timestamps (begin and end of the anomaly) and a label identifying the anomalous device. I think if a device is annotated as anomalous at two non-overlapping time period this should be two separate anomalies
Okay, it sounds like my concern is bunk. Let me rephrase, though, just to be clear. I think what you're saying is, if there is an anomaly A that starts at t_1 and ends at t_2, then anomaly A will always exist. That is, no behaviors at t_3 > t_2 will make the earlier anomaly A appear "non-anomalous." Or, in crude terms, "once an anomaly, always an anomaly."
yes it is, anomaly forever :-)
done! the graph are now generated based on the url parameters
Awesome. @romain-fontugne we need the files put somewhere so we could run our own tests. How big in the file you're using to run your tests? Can we drop it somewhere or can you put it somewhere i could download it from?
Charts need to be generated from url parameters so that we could include the url in an email that generates the anomaly report for a particular file.