Is there a function to do frequency (e.g., Fourier) analysis on time series data? I see periodic (every hour) activity spikes on time series data and would like to be able to detect the spikes automatically and alert on them.
One way to do this, given that I'm interested in correlating now vs. 1 hour ago, is as follows:
signal = ts(data, host=somehost)
lagsignal = lag(60m, ts(data, host=somehost))
Then look at mcorr(10m, ${signal}, ${lagsignal}).
