Linux disk IO check. Reads /sys filesystem directly to query tps, read, write, avg. request size, avg. queue size and avg. wait time.
Capable of graphing and alerting on the disk queue length as seen in `iostat -x'.
INSTALL
Example:
wget https://raw.github.com/mclarkson/check_diskstat/master/check_diskstat.sh chmod +x check_diskstat.sh mv check_diskstat.sh /usr/lib64/nagios/plugins/
EXAMPLE OUTPUT
$ ./check_diskstat.sh -d sda -W 10 -C 20 summary: 3 io/s, read 32 sectors (0kB/s), write 4216 sectors (43kB/s), queue size 0 in 48 seconds | tps=3io/s;;; read=341b/s;;; write=44970b/s;;; avgrq-sz=24;;; avgqu-sz=0;10;20; await=43ms;;;
MANUAL PAGE
./check_diskstat.sh -d DEVICE [ -w tps,read,write -c tps,read,write ]
| [ -W qlen -C qlen ] | -h
This plug-in is used to be alerted when maximum hard drive io/s, sectors
read|write/s or average queue length is reached.
-d DEVICE DEVICE must be without /dev (ex: -d sda).
To specify a LVM logical volume use:
volgroup/logvol.
To specify symlink from /dev/disk/ use full path, ex:
/dev/disk/by-id/scsi-35000c50035006fb3
-w/c TPS,READ,WRITE TPS means transfer per seconds (aka IO/s)
READ and WRITE are in sectors per seconds
-W/C NUM Use average queue length thresholds instead..
-b Brief output.
-s silent output: no warnings or critials are issued
Performance data for graphing is supplied for tps, read, write, avgrq-sz,
avgqu-sz and await (see iostat man page for details).
Example: Tps, read and write thresholds:
./check_diskstat.sh -d sda -w 200,100000,100000 -c 300,200000,200000
Example: Average queue length threshold:
./check_diskstat.sh -d sda -W 50 -C 100