4 ms·
Another interesting generalization of the DFT is the Lomb-Scargle transformation, which doesn't require a fixed measurement interval in the time domain. It is
by perlgeek 3y ago
Another interesting generalization of the DFT is the Lomb-Scargle transformation, which doesn't require a fixed measurement interval in the time domain.
It is often used to determine the frequency of a period signal when you don't have fixed measurement interval, like in astrophysics.
https://iopscience.iop.org/article/10.3847/1538-4365/aab766 https://iopscience.iop.org/article/10.3847/1538-4365/aab766 has a general introduction, and https://docs.astropy.org/en/stable/timeseries/lombscargle.html https://docs.astropy.org/en/stable/timeseries/lombscargle.ht... is a nice intro to using it in the astropy library (in Python)
- hinkley 3y agoI’ve been thinking a lot lately about how useful it might be to represent Prometheus data in a frequency domain, to visualize weekly, daily, and yearly access patterns in capacity planning. Autoscaling can avoid brownouts but it can’t tell you what your annual budget should be. Or why. But Prometheus data isn’t really a sampling interval. Even if each machine in you cluster is reporting on an interval, they aren’t synchronized.