Additional scripts for Zabbix agent on Linux to discover and monitor several services
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 

278 lines
8.8 KiB

#!/usr/bin/perl -w
use strict;
use warnings;
use Config::Simple '-strict';
use Getopt::Long;
use File::Basename;
# Output file
my $output = undef;
# When a threshold can be automatically detected,
# you may want to be notified before it's reached, so you can
# set a margin which will be substracted from the real threshold
my $temp_margin = '20';
my $temp_hd_margin = '10';
my $pwr_margin = '200';
#This value will be substracted from the higher threshold to define the low one
#so you can have hysteresis to prevent flip-flop
my $temp_hyst = '10';
my $temp_hd_hyst = '5';
my $pwr_hyst = '200';
GetOptions(
"output=s" => \$output,
"temp-margin=i" => \$temp_margin,
"temp-hd-margin=i" => \$temp_hd_margin,
"pwr-margin=i" => \$pwr_margin,
"temp-hyst=i" => \$temp_hyst,
"temp-hd-hyst=i" => \$temp_hd_hyst
);
sub usage(){
print<<"_EOF";
Usage: $0 --output=/etc/zabbix/sensors.ini
_EOF
}
unless ($output){
usage();
exit 1;
}
# Path
my $ipmitool = '/usr/bin/ipmitool';
my $smartctl = '/usr/sbin/smartctl';
# Default threshold if not detected
my $def_temp_thres_high = '50';
my $def_fan_thres_high = '1000';
my $def_fan_thres_low = '700';
my $def_pwr_thres_high = '1000';
my $cfg = new Config::Simple(syntax => 'ini');
my $sensors = {};
# Try to detect IPMI sensors
if (-x $ipmitool){
# First check for temperature sensors
my @lines = qx($ipmitool sdr type Temperature);
if ($? == 0){
SENSOR: foreach my $l (@lines){
chomp $l;
# Looks like
# Inlet Temp | 04h | ok | 7.1 | 25 degrees C
if ($l !~ m/^(\w+[\s\w]+?\w+)\s*\|.*\|\s*([\w\.\s]+)\s*\|.*\|\s*([\-\w\.\s]+)$/){
next SENSOR;
}
my $name = $1;
my $sensor = {};
my @details = qx($ipmitool sdr get '$name');
if ($? != 0){
print "Couldn't get detail for sensor $name\n";
next SENSOR;
}
foreach my $d (@details){
chomp $d;
if ($d =~ m/^\s*Sensor\sReading\s*:\s*(\w+)/){
my $val = $1;
if ($val !~ m/^\d+$/){
print "Skipping sensor $name, couldn't parse its value: $val\n";
next SENSOR;
}
}
elsif ($d =~ m/^\s*Upper\scritical\s*:\s*(\d+(\.\d+))/){
$sensor->{threshold_high} = $1-$temp_margin;
}
elsif ($d =~ m/^\s*Upper\snon\-critical\s*:\s*(\d+(\.\d+))/){
$sensor->{threshold_low} = $1-$temp_margin;
}
}
$sensor->{threshold_high} ||= $def_temp_thres_high;
$sensor->{threshold_low} ||= $def_temp_thres_high-$temp_hyst;
$sensor->{threshold_high} =~ s/\.0+$//;
$sensor->{threshold_low} =~ s/\.0+$//;
$sensor->{description} = $name;
$sensor->{type} = 'temp';
$sensor->{unit} = '°C';
$sensor->{cmd} = "$ipmitool sdr get '$name' | grep 'Sensor Reading' | awk '{print \$4}'";
my $id = lc $name;
$id =~ s/\s/_/g;
$sensors->{$id} = $sensor;
print "Found a temperature sensor using IPMI: $name\n";
}
}
# Now check for Fan, nearly the same as Temp, but
# * We try to detect the unit
# * threshold handling is not the same
@lines = qx($ipmitool sdr type Fan);
if ($? == 0){
SENSOR: foreach my $l (@lines){
chomp $l;
$l =~ m/^(\w+[\s\w]+?\w+)\s*\|.*\|\s*([\w\.\s]+)\s*\|.*\|\s*([\-\w\.\s]+)$/;
my $name = $1;
my $value = my $val = $3;
my $sensor = {};
my @details = qx($ipmitool sdr get '$name');
if ($? != 0){
print "Couldn't get detail for sensor $name\n";
next SENSOR;
}
foreach my $d (@details){
chomp $d;
if ($d =~ m/^\s*Sensor\sReading\s*:\s*(\w+)/){
$val = $1;
if ($val !~ m/^\d+$/){
print "Skipping sensor $name, couldn't parse its value: $val\n";
next SENSOR;
}
}
elsif ($d =~ m/^\s*Lower\scritical\s*:\s*(\d+(\.\d+))/){
$sensor->{threshold_low} = $1-$temp_margin;
}
elsif ($d =~ m/^\s*Lower\snon\-critical\s*:\s*(\d+(\.\d+))/){
$sensor->{threshold_high} = $1-$temp_margin;
}
}
$sensor->{threshold_high} ||= $def_fan_thres_high;
$sensor->{threshold_low} ||= $def_fan_thres_high-$temp_hyst;
$sensor->{threshold_high} =~ s/\.0+$//;
$sensor->{threshold_low} =~ s/\.0+$//;
$sensor->{description} = $name;
$sensor->{type} = 'fan';
$sensor->{unit} = ($value =~ m/percent|%/ || $val < 100) ? '%' : 'rpm';
$sensor->{cmd} = "$ipmitool sdr get '$name' | grep 'Sensor Reading' | awk '{print \$4}'";
my $id = lc $name;
$id =~ s/\s/_/g;
$sensors->{$id} = $sensor;
print "Found a fan sensor using IPMI: $name\n";
}
}
# Now look for power information
@lines = qx($ipmitool sdr type 'Current');
if ($? == 0){
SENSOR: foreach my $l (@lines){
chomp $l;
$l =~ m/^(\w+[\s\w]+?\w+)\s*\|.*\|\s*([\w\.\s]+)\s*\|.*\|\s*([\-\w\.\s]+)$/;
my $name = $1;
my $val = $3;
my $sensor = {};
if ($name =~ m/Power|Pwr|Consumption/i || $val =~ m/W(att)?/i){
my @details = qx($ipmitool sdr get '$name');
if ($? != 0){
print "Couldn't get detail for sensor $name\n";
next SENSOR;
}
foreach my $d (@details){
chomp $d;
if ($d =~ m/^\s*Sensor\sReading\s*:\s*(\w+)/){
$val = $1;
if ($val !~ m/^\d+$/){
print "Skipping sensor $name, couldn't parse its value: $val\n";
next SENSOR;
}
}
elsif ($d =~ m/^\s*Upper\scritical\s*:\s*(\d+(\.\d+))/){
$sensor->{threshold_high} = $1-$pwr_margin;
}
elsif ($d =~ m/^\s*Upper\snon\-critical\s*:\s*(\d+(\.\d+))/){
$sensor->{threshold_low} = $1-$pwr_margin;
}
}
$sensor->{threshold_high} ||= $def_pwr_thres_high;
$sensor->{threshold_low} ||= $def_pwr_thres_high-$pwr_hyst;
$sensor->{threshold_high} =~ s/\.0+$//;
$sensor->{threshold_low} =~ s/\.0+$//;
$sensor->{description} = $name;
$sensor->{type} = 'power';
$sensor->{unit} = 'Watt';
$sensor->{cmd} = "$ipmitool sdr get '$name' | grep 'Sensor Reading' | awk '{print \$4}'";
my $id = lc $name;
$id =~ s/\s/_/g;
$sensors->{$id} = $sensor;
print "Found a power sensor using IPMI: $name\n";
}
}
}
}
# Now, try to detect smart capable HDD
if (-x $smartctl){
# This is copied from disco_smart_sudo, I should create a module to
# consolidate this
opendir(my $dh, "/sys/block") or die "Couldn't open /sys/block: $!";
my @blocks = grep { $_ !~ m/^\./ } readdir($dh);
closedir($dh);
foreach my $block (@blocks){
my $removable = 0;
my $size = 1;
# Skip block we already know they won't support SMART
next if ($block =~ m/^(ram|loop|md|dm\-)\d+/);
if ( -e "/sys/block/$block/removable"){
open REMOVABLE, "/sys/block/$block/removable";
$removable = join "", <REMOVABLE>;
close REMOVABLE;
chomp($removable);
next if ($removable eq '1');
}
if ( -e "/sys/block/$block/size"){
open SIZE, "/sys/block/$block/size";
$size = join "", <SIZE>;
close SIZE;
chomp($size);
next if ($size eq '0');
}
my @lines = qx($smartctl -A /dev/$block);
next if ($? != 0);
foreach my $l (@lines){
if ($l =~ /Temperature_Celsius/){
my $sensor = {
description => "$block temperature",
threshold_low => $def_temp_thres_high-$temp_hd_hyst,
threshold_high => $def_temp_thres_high,
type => 'temp',
unit => '°C',
cmd => "$smartctl -A /dev/$block | grep Temperature_Celsius | awk '{print \$10}'"
};
$sensors->{$block} = $sensor;
print "Found a temperature sensor using smartctl: $block\n";
last;
}
}
}
# Some LSI based hardware RAID controller can report HDD temp
if (-e '/dev/megaraid_sas_ioctl_node'){
# Only check for the firsts 26 drives
foreach my $i (0..25){
my @res = qx($smartctl -d megaraid,$i -A /dev/sda);
next if ($? != 0);
foreach my $l (@res){
if ($l =~ m/Drive\sTrip\sTemperature:\s+(\d+)\s/){
my $sensor = {
description => "Temperature for disk No $i on sda",
type => 'temp',
threshold_high => $1-$temp_hd_margin,
threshold_low => $1-$temp_hd_margin-$temp_hd_hyst,
unit => '°C',
cmd => "$smartctl -A -d megaraid,$i /dev/sda | grep 'Current Drive Temperature' | awk '{print \$4}'"
};
$sensors->{'sda-' . $i} = $sensor;
print "Found a temperature sensor using smartctl (megaraid): sda-$i\n";
}
}
}
}
}
# TODO: add support for lm sensors, but its ouput is harder to parse
foreach my $s (keys %$sensors){
$cfg->set_block($s, $sensors->{$s});
}
$cfg->write($output);