Module: check_mk
Branch: master
Commit: 51dbd05f5472478a02e26f155e48481c178de416
URL:
http://git.mathias-kettner.de/git/?p=check_mk.git;a=commit;h=51dbd05f547247…
Author: Bernd Stroessenreuther <bs(a)mathias-kettner.de>
Date: Fri Jul 26 12:00:39 2013 +0200
initial version of check manpage for statgrab_*, added section agentless in check catalog
---
checkman/statgrab_disk | 28 ++++++++++++++++-------
checkman/statgrab_mem | 52 ++++++++++++++++++++++++++++++++++--------
checkman/statgrab_net.ctr | 26 ++++++++++++---------
checkman/statgrab_net.link | 26 +++++++++++++--------
checkman/statgrab_net.params | 2 +-
doc/drafts/LIESMICH.Katalog | 2 ++
6 files changed, 97 insertions(+), 39 deletions(-)
diff --git a/checkman/statgrab_disk b/checkman/statgrab_disk
index 968e27f..c4e0c33 100644
--- a/checkman/statgrab_disk
+++ b/checkman/statgrab_disk
@@ -4,16 +4,28 @@ catalog: os/kernel
license: GPL
distribution: check_mk
description:
- This check measures the CPU utilization during the last check
- cycle. A level can be set only on {wait} (disk wait).
+ Reports Disk IO for read and Disk IO for write in Bytes/s during the last check
+ cycle.
- BERND: Hinweis auf statgrab-Plugin (HP) bzw. statgrab-Befehl (beide).
+ The check always returns {OK} state.
+
+ Make sure you have the statgrab binary available on the monitored
+ machine. Under HP-UX you also need to activate the hpux_statgrab agent plugin.
+
+item:
+ {read} for Disk IO read and
+ {write} for Disk IO write
+
+examples:
+ # adding one service for Disk IO read and one for Disk IO write
+ checks += [
+ ('statgrab', 'statgrab_disk', 'read', None),
+ ('statgrab', 'statgrab_disk', 'write', None),
+ ]
perfdata:
- Three variables: The percentage of time spent in {user}, {system} and {wait}
- (disk wait). The idle time can be computed by substracting those three values
- from 100%. The user value contains also {nice}.
+ One value per Check: Disk IO in Bytes/s
inventory:
- One check per host is created if the agent contains a {statgrab_cpu}
- section containing cpu information.
+ Finds two checks on every machine with statgrab: Disk IO read and
+ Disk IO write
diff --git a/checkman/statgrab_mem b/checkman/statgrab_mem
index 4224e7d..cd486ff 100644
--- a/checkman/statgrab_mem
+++ b/checkman/statgrab_mem
@@ -1,19 +1,53 @@
-title: Memory (RAM) and Swap usage
+title: Usage of physical and virtual RAM
agents: hpux, solaris
catalog: os/kernel
license: GPL
distribution: check_mk
description:
- This check measures the CPU utilization during the last check
- cycle. A level can be set only on {wait} (disk wait).
+ This check measures the current usage of physical RAM and
+ virtual memory used by processes. You can define a warning
+ and critical level for the usage of virtual memory,
+ {not} for the usage of RAM.
- BERND: Hinweis auf statgrab-Plugin (HP) bzw. statgrab-Befehl (beide).
+ Check_mk lets you define levels in percentage of
+ the physically installed RAM or as absolute
+ values in MB. The default levels
+ are at 150% and 200%. That means that this check
+ gets critical if the memory used by processes
+ is twice the size of your RAM.
+
+ See check manpage for {mem.used} too.
+
+ Make sure you have the statgrab binary available on the monitored
+ machine. Under HP-UX you also need to activate the hpux_statgrab agent plugin.
+
+examples:
+ # Change default levels from 150%/200% to 100%/150%:
+ memused_default_levels = (100.0, 150.0)
+ checks += [
+ # make explicit check for hosts with tag "solaris"
+ ( ["solaris"], ALL_HOSTS, "statgrab_mem", None, (80.0, 100.0) ),
+ # use absolute levels at 8GB / 12GB for some other hosts
+ ( ["host12","host34"], "statgrab_mem", None, (8192,
12288) )
+ ]
perfdata:
- Three variables: The percentage of time spent in {user}, {system} and {wait}
- (disk wait). The idle time can be computed by substracting those three values
- from 100%. The user value contains also {nice}.
+ Two values: The RAM used in MB (without Caches) and the swap space used
+ in MB.
inventory:
- One check per host is created if the agent contains a {statgrab_cpu}
- section containing cpu information.
+ Finds one service on every machine with statgrab: The memory used.
+
+[parameters]
+warning (int or float): the percentage of virtual memory used
+ by processes at which {WARN} state is triggered. If the
+ level is defined as an integer value then it is interpreted
+ as an absolute value in megabytes.
+critical (int or float): the percentage or absolute value
+ at which {CRIT} state is triggered
+
+[configuration]
+memused_default_levels (float, float): Levels used by
+ all checks that are created by inventory. If not set they
+ have defaults of (150.0, 200.0)
+
diff --git a/checkman/statgrab_net.ctr b/checkman/statgrab_net.ctr
index c2158d7..3b938f7 100644
--- a/checkman/statgrab_net.ctr
+++ b/checkman/statgrab_net.ctr
@@ -1,20 +1,24 @@
-title: Network Interfaces: Counters
+title: Network Interfaces: Traffic
agents: hpux, solaris
-catalog: os/kernel
+catalog: os/network
license: GPL
distribution: check_mk
description:
- This check measures the CPU utilization during the last check
- cycle. A level can be set only on {wait} (disk wait).
+ Reports the network traffic for an interface during the last check
+ cycle for direction IN and OUT.
- BERND: Hinweis auf statgrab-Plugin (HP) bzw. statgrab-Befehl (beide).
- Title to be improved
+ Make sure you have the statgrab binary available on the monitored
+ machine. Under HP-UX you also need to activate the hpux_statgrab agent plugin.
+
+item:
+ The interface name as defined in the operating system.
perfdata:
- Three variables: The percentage of time spent in {user}, {system} and {wait}
- (disk wait). The idle time can be computed by substracting those three values
- from 100%. The user value contains also {nice}.
+ Seven values are delivered, the differences of the counters during the last check
+ cycle for {rx_bytes}, {tx_bytes}, {rx_packets}, {tx_packets}, {rx_errors},
+ {tx_errors}, {tx_collisions}
inventory:
- One check per host is created if the agent contains a {statgrab_cpu}
- section containing cpu information.
+ Creates one check for each network interface.
+
+
diff --git a/checkman/statgrab_net.link b/checkman/statgrab_net.link
index 84bf0a5..3e7d618 100644
--- a/checkman/statgrab_net.link
+++ b/checkman/statgrab_net.link
@@ -1,19 +1,25 @@
title: Network Interfaces: Link status
agents: hpux, solaris
-catalog: os/kernel
+catalog: os/network
license: GPL
distribution: check_mk
description:
- This check measures the CPU utilization during the last check
- cycle. A level can be set only on {wait} (disk wait).
+ BERND, FIXME: where does it get target state from?
- BERND: Hinweis auf statgrab-Plugin (HP) bzw. statgrab-Befehl (beide).
+ Checks the link state for an interface and compares it to it's target state.
-perfdata:
- Three variables: The percentage of time spent in {user}, {system} and {wait}
- (disk wait). The idle time can be computed by substracting those three values
- from 100%. The user value contains also {nice}.
+ Returns {OK} if link state is equal to the target state.
+
+ Returns {WARN} if the link is UP and the target state is UNUSED.
+
+ Returns {CRIT} if the state is NO LINK and the target state is UP.
+
+ Make sure you have the statgrab binary available on the monitored
+ machine. Under HP-UX you also need to activate the hpux_statgrab agent plugin.
+
+item:
+ The interface name as defined in the operating system.
inventory:
- One check per host is created if the agent contains a {statgrab_cpu}
- section containing cpu information.
+ Creates one check for each network interface.
+
diff --git a/checkman/statgrab_net.params b/checkman/statgrab_net.params
index 799335d..eaad10f 100644
--- a/checkman/statgrab_net.params
+++ b/checkman/statgrab_net.params
@@ -1,6 +1,6 @@
title: Network Interfaces: Speed and Duplex
agents: hpux, solaris
-catalog: os/kernel
+catalog: os/network
license: GPL
distribution: check_mk
description:
diff --git a/doc/drafts/LIESMICH.Katalog b/doc/drafts/LIESMICH.Katalog
index 86a6816..bf39c34 100644
--- a/doc/drafts/LIESMICH.Katalog
+++ b/doc/drafts/LIESMICH.Katalog
@@ -52,3 +52,5 @@ hw
SNMP-Agent auf Betriebssystem für HW
generic (mrpe, local, ...)
+
+agentless "Networking checks without agent"