This check is inspired by this article.
It’s common to check for swap usage above a threshold, but even if you have a small quantity of memory swapped, it’s actually the rate it’s swapped in/out that can impact performance, not the quantity. This is a much more direct check for that state.
A check should be instant. In order to avoid waiting several seconds for vmstat to compute the deritative of pswpin/pswpout, this check stores the current state in a file. The result is instant (no hang).
This check comes with performance data (used to generate a graph from a check).
make
For now, it's only built/tested on Linux.
$ ./check_swap_rate -h
Usage: ./check_swap_rate [-i INTERVAL] [-w W_THRESHOLD] [-c C_THRESHOLD] [-h] [-V]
-i INTERVAL interval in seconds (default 60)
-c C_THRESHOLD number of pages for critical threshold (default 60)
-w W_THRESHOLD number of pages for warning threshold (default 30)
-h this help
-V version
Examples:
./check_swap_rate -i 2 -c 2 -w 1 # critical: 1 page/s warning: 0.5 page/s
./check_swap_rate -i 600 -c 1200 -w 300 # critical: 2 page/s warning: 0.5 page/s
Note: INTERVAL is only a mean to provide the rate, which in the end is a number of pages per sec.