measure_overload: Measure overload
In atusy/pingAnalysis: Analyze Ping Log Files

Description Usage Arguments Value Examples

View source: R/measure_overload.R

Summarize the overload conditions when rolling mean of ping exceeds the limit. If a window contains timeout, it is not considered as overload regardless of any finite values within the window. Note that no-response is discarded instead of being treated as timeout by measure_timeout().

1	measure_overload(log_df, m = 2L, t = 10, N = 1L)

`log_df`	A `data.frame` that records the log.
`m`	A window size of rolling mean of ping time.
`t`	Upper limit of rolling mean of ping time.
`N`	Threshold of continuous errors to be considered as fatal.

A data frame with following columns:

address: (chr) IPv4 addresses that suffered timeout.
start: (dttm) When server started to timeout.
end: (dttm) When server resumed or NA if not resumed.
n_timeout: (int) Number of timeout before resume.

data.frame

log_df <- data.frame(
  timestamp = lubridate::ymd_hms(20200101000000 + seq(6)),
  address = "192.168.1.1/24",
  ping = c(1, 10, NA, 10, 10, NA)
)
measure_overload(log_df, m = 1L, t = 10)
measure_overload(log_df, m = 2L, t = 10)