/ 
Monitoring-Sneck-Checks-0.0.2
/ Monitoring::Sneck::Checks::pressure

NAME

Monitoring::Sneck::Checks::pressure - checks Linux pressure stall information

VERSION

Version 0.0.1

SYNOPSIS

pressure|perl:pressure
memory_pressure|perl:pressure -r memory -k full -w 5 -c 20
io_pressure|perl:pressure -r io --io-warning 30 --io-critical 60 -a 300

DESCRIPTION

Checks the percent of time tasks were stalled waiting on CPU, memory, or IO, from /proc/pressure. This tells being short of each apart in a way load does not.

Each resource has some, the time at least one task was stalled, and full, the time all were at once, which is the worse. Each is averaged over 10, 60, and 300 seconds. full for CPU is always 0 for the system as a whole, so it is left out.

Only Linux 4.20 and later have this, and only if not turned off with psi=0. Others are UNKNOWN.

OPTIONS

- -r, --resource :: cpu, memory, or io. May be given more than once. Default :: all three

- -k, --kind :: some or full. Default :: some

- -a, --average :: 10, 60, or 300 seconds. Default :: 60

- -w, --warning :: Range of percent stalled. Default :: 20

- -c, --critical :: Range of percent stalled. Default :: 50

- --cpu-warning, --cpu-critical, --memory-warning, --memory-critical, --io-warning, --io-critical :: The same, for only that resource.

METHODS

check

Does the check.