Ranges for performance data

c-holtermann commented 9 years ago

I wrote a little nagios_plugin to check btrfs

#!/usr/bin/python
# Von https://pythonhosted.org/NagAconda/plugin.html
from NagAconda import Plugin
import os
import subprocess
from pyparsing import *

btrfs_check = Plugin("Plugin to show disk usage of btrfs.", "0.1")
btrfs_check.add_option('m', 'mountpoint', 'mountpoint for btrfs',
    required=True)

btrfs_check.enable_status('warning')
btrfs_check.enable_status('critical')
btrfs_check.start()

btrfs_output = subprocess.check_output(["btrfs", "fi", "df", btrfs_check.options.mountpoint])

# DEBUG: 
# print btrfs_output

# PyParsing definitions
#
# Output is something like:
# Data, RAID1: total=222.72GB, used=197.44GB
# System, RAID1: total=64.00MB, used=40.00KB
# System: total=4.00MB, used=0.00
# Metadata, RAID1: total=10.00GB, used=5.40GB

# Parse Byte values with units
byteDefs=["B", "KB", "MB", "GB", "TB", "PB", "EB"]
byteDef = oneOf(byteDefs)
fnumber = Word(nums+'.').setParseAction(lambda t: float(t[0]))
bnumber = Group(fnumber('num') + Optional(byteDef('unit'), default="B"))

# Definitions for each row
row_data = "Data, " + "RAID1: " + "total=" + bnumber('bytesize_total') + "," + "used=" + bnumber('bytesize_used')
row_system1 = "System, " + "RAID1: " + "total=" + bnumber('bytesize_total') + "," + "used=" + bnumber('bytesize_used')
row_system2 = "System: " + "total=" + bnumber('bytesize_total') + "," + "used=" + bnumber('bytesize_used')
row_metadata = "Metadata, " + "RAID1: " + "total=" + bnumber('bytesize_total') + "," + "used=" + bnumber('bytesize_used')

# The whole parsing term
btrfs_output_parser = row_data('data') + row_system1('system1') + row_system2('system2') + row_metadata('metadata')

# Parse btrfs output
btrfs_output_parsed = btrfs_output_parser.parseString(btrfs_output)

# calculate values
byteUnits=[("B", 1), ("KB", 1000), ("MB", 1000000), ("GB", 1000000000), ("TB", 1000000000000), ("PB", 1000000000000000), ("EB", 1000000000000000000)]
def calc_bnumber(bnum):
        """Get float number of bytes for something like 2GB"""
        for byteSize in byteUnits:
                if byteSize[0] == bnum.unit:
                        return byteSize[1] * bnum.num
        return None

data_bytes_used_float = calc_bnumber(btrfs_output_parsed.data.bytesize_used)
data_bytes_size_float = calc_bnumber(btrfs_output_parsed.data.bytesize_total)
data_usage_percentage_float = data_bytes_used_float / data_bytes_size_float * 100.0

system1_bytes_used_float = calc_bnumber(btrfs_output_parsed.system1.bytesize_used)
system1_bytes_size_float = calc_bnumber(btrfs_output_parsed.system1.bytesize_total)

system2_bytes_used_float = calc_bnumber(btrfs_output_parsed.system2.bytesize_used)
system2_bytes_size_float = calc_bnumber(btrfs_output_parsed.system2.bytesize_total)

metadata_bytes_used_float = calc_bnumber(btrfs_output_parsed.metadata.bytesize_used)

# DEBUG:
# print btrfs_output_parsed.dump()

# set nagios output
btrfs_check.set_range('warning', 100000000000000000000000, range_num=2)
btrfs_check.set_range('critical', 200000000000000000000000, range_num=2)

btrfs_check.set_value("data_used", btrfs_output_parsed.data.bytesize_used.num, scale=btrfs_output_parsed.data.bytesize_used.unit, threshold=2)
btrfs_check.set_value("data_total", btrfs_output_parsed.data.bytesize_total.num, scale=btrfs_output_parsed.data.bytesize_total.unit, threshold=2)
btrfs_check.set_value("data_ratio", data_usage_percentage_float, scale="%", threshold=1)

btrfs_check.set_value("system1_used", btrfs_output_parsed.system1.bytesize_used.num, scale=btrfs_output_parsed.system1.bytesize_used.unit, threshold=2)
btrfs_check.set_value("system1_total", btrfs_output_parsed.system1.bytesize_total.num, scale=btrfs_output_parsed.system1.bytesize_total.unit, threshold=2)

btrfs_check.set_value("system2_used", btrfs_output_parsed.system2.bytesize_used.num, scale=btrfs_output_parsed.system2.bytesize_used.unit, threshold=2)
btrfs_check.set_value("system2_total", btrfs_output_parsed.system2.bytesize_total.num, scale=btrfs_output_parsed.system2.bytesize_total.unit, threshold=2)

btrfs_check.set_value("metadata_used", btrfs_output_parsed.metadata.bytesize_used.num, scale=btrfs_output_parsed.metadata.bytesize_used.unit, threshold=2)
btrfs_check.set_value("metadata_total", btrfs_output_parsed.metadata.bytesize_total.num, scale=btrfs_output_parsed.metadata.bytesize_total.unit, threshold=2)

btrfs_check.finish()

I'd like to only check the data_ratio with the standard warning and critical values. This actually works when adding a second range that all the other performance datas are checked against. But the output doesn't care about this.

When being called by

nagaconda_btrfs_df.py -m /mounts/BackupRAID -w0.9 -c0.95

Status Ok|metadata_used=5400000000.0B;0.9;0.95;; data_total=222.72GB;0.9;0.95;; data_ratio=0.886494252874;0.9;0.95;; data_used=197.44GB;0.9;0.95;; system2_used=0.0B;0.9;0.95;; metadata_total=10000000000.0B;0.9;0.95;; system1_used=40.0KB;0.9;0.95;; system2_total=4000000.0B;0.9;0.95;; system1_total=64.0MB;0.9;0.95;;

I'd expect the 2nd threshold not only in the tests but also in the output

c-holtermann commented 9 years ago

This plugin resides at https://github.com/c-holtermann/check_btrfs

c-holtermann commented 9 years ago

It would be nice if the documentation would be more clear about the difference between the main parameter that triggers the levels OK, WARN, CRIT and the other performance parameters

bonesmoses / NagAconda

Ranges for performance data #7