Discussion thread for official Zabbix Template for Linux

[jeff16384]

Hi, some of our Linux servers are configured to heavily swap unused memory do get free memory for disk cache.

as a result I often see memory situations like this:


so free swap is less than 25% but lot of memory is available. so swap-in of all data is always possible. which means we often have Alerts about High swap usage by zabbix. I suggest this small patch:

--- linux-template.orig 2023-08-24 16:54:31.439127713 +0000
+++ linux-template.new 2023-08-24 16:54:53.458292768 +0000
@@ -479,6 +479,17 @@
value: memory
- tag: component
value: storage
+ - uuid: 2e4e665e1b664d889ad690692e031aba
+ name: 'Linux: Used swap space'
+ key: 'system.swap.size[,used]'
+ history: 7d
+ units: B
+ description: 'The used space of the swap volume/file expressed in bytes.'
+ tags:
+ - tag: component
+ value: memory
+ - tag: component
+ value: storage
- uuid: fc6898682f904c57ba28bcb3748b237f
name: 'Linux: System description'
key: system.uname
@@ -1921,12 +1932,14 @@
- tag: scope
value: performance
- uuid: f49fc01c5ac945d59455563bdbcc6469
- expression: 'max(/Linux by Zabbix agent/system.swap.size[,pfree],5m)<{$SWAP.PFREE.MIN.WARN} and last(/Linux by Zabbix agent/system.swap.size[,total])>0'
+ expression: 'max(/Linux by Zabbix agent/system.swap.size[,pfree],5m)<{$SWAP.PFREE.MIN.WARN} and last(/Linux by Zabbix agent/system.swap.size[,total])>0 and last(/Linux by Zabbix agent/system.swap.size[,used]) > last(/Linux by Zabbix agent/vm.memory.size[available])'
name: 'Linux: High swap space usage'
- event_name: 'Linux: High swap space usage (less than {$SWAP.PFREE.MIN.WARN}% free)'
+ event_name: 'Linux: High swap space usage (less than {$SWAP.PFREE.MIN.WARN}% free) and available memory lower than used swap space'
opdata: 'Free: {ITEM.LASTVALUE1}, total: {ITEM.LASTVALUE2}'
priority: WARNING
- description: 'If there is no swap configured, this trigger is ignored.'
+ description: |
+ If there is no swap configured, this trigger is ignored.
+ Do not trigger as long as enough memory is available
dependencies:
- name: 'Linux: High memory utilization'
expression: 'min(/Linux by Zabbix agent/vm.memory.utilization,5m)>{$MEMORY.UTIL.MAX}'


explanation:
1. Polling also Used Swapspace
2. only Trigger "High Swap usage" if available memory is less than used swap.

BR. Jörg

[pfzim]

Hi,
I noticed that in the Linux templates, the condition for absolute values was removed, leaving only percentages.
I believe this change is a mistake because for large-capacity disks, 20% is still a significant amount of free space for a trigger to activate.
For example, on a 2 TB disk, the trigger activates at 20%, which is 400 GB of free space. Even SCOM uses a combined trigger: https://kevinholman.com/2019/07/09/h...works-in-scom/

Triggers:

• FS [{#FSNAME}]: Space is critically low
• FS [{#FSNAME}]: Space is low

The commit with the change: https://git.zabbix.com/projects/ZBX/...31d7b5dfac4ccd

I suggest reverting it to how it was:

last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},pused])>{$VFS.FS.PUSED.MAX.CRIT:"{#FSNAME}"} and
(last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},total])-last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},used]))<{$VFS.FS.FREE.MIN.CRIT:"{#FSNAME}"}

[pfzim]

Event name:

Linux: FS [{#FSNAME}]: Space is low (used > {$VFS.FS.FREE.MIN.WARN:"{#FSNAME}"} and {$VFS.FS.PUSED.MAX.WARN:"{#FSNAME}"}%, free {{?(last(//vfs.fs.dependent.size[{#FSNAME},total])-last(//vfs.fs.dependent.size[{#FSNAME},used]))/1024/1024/1024}.fmtnum(1)}GB, total {{?last(//vfs.fs.dependent.size[{#FSNAME},total])/1024/1024/1024}.fmtnum(1)}GB)

Linux: FS [{#FSNAME}]: Space is critically low (used > {$VFS.FS.FREE.MIN.CRIT:"{#FSNAME}"} and {$VFS.FS.PUSED.MAX.CRIT:"{#FSNAME}"}%, free {{?(last(//vfs.fs.dependent.size[{#FSNAME},total])-last(//vfs.fs.dependent.size[{#FSNAME},used]))/1024/1024/1024}.fmtnum(1)}GB, total {{?last(//vfs.fs.dependent.size[{#FSNAME},total])/1024/1024/1024}.fmtnum(1)}GB)

Operational data:

Space used: {ITEM.LASTVALUE3} ({{ITEM.LASTVALUE1}.fmtnum(1)}%), total: {ITEM.LASTVALUE2})}

Expression:

last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},pused])>{$VFS.FS.PUSED.MAX.WARN:"{#FSNAME}"} and
(last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},total])-last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},used]))<{$VFS.FS.FREE.MIN.WARN:"{#FSNAME}"}

last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},pused])>{$VFS.FS.PUSED.MAX.CRIT:"{#FSNAME}"} and
(last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},total])-last(/Linux by Zabbix agent active/vfs.fs.dependent.size[{#FSNAME},used]))<{$VFS.FS.FREE.MIN.CRIT:"{#FSNAME}"}​

[dharam]

Anyone using File system monitoring for Linux host and can provide me high level steps to begin with. I have given task to explore this option.
Currently we have script based from tool called HP ovo and we need to replace it with zabbix

[dharam]

Hi,
I have downloaded the template - Linux by Zabbix agent active and applied to Zabbix app and proxy servers (both are on linux) , I need to do any configuration to discover any filesystem on this? I dont see any file system in latest data.
Any help on this?

[PavelZ]

This is a pretty popular and reliable template. Nothing breaks there.
Maybe you show versions of the agent, template, and also check the values of the host macros ? (this affects filesystems autodiscovery)

[dharam]

Hi,
Please find below details :
Zabbix 7.0.12
Agent version : 7.0.12
fstype: ext4 xfs (how to discover other file system e.g tmpfs,other mount fs)

Item : Get FileSystems - shows no data (when tried to run discovery shows , shows attached error)
Template : Linux by Zabbix agent active (Zabbix 7.0 or higher)

Detailed snapshot.

[PavelZ]

Item : Get FileSystems - shows no data (when tried to run discovery shows , shows attached error)

This function does not work only because the agent kind is active
Can you debug querying of information using zabbix_get program separately?

[dharam]

sorry but can you give me how to debug this as I am not aware?

[ralfbergs]

Hi.

I would suggest that this template be extended as follows:

• Include an (optional) protocol scheme macro so that you can change from "http://" URLs to "https://" URLs.
• Include (optional) basic authentication credentials "login" and "password".

I don't think these features are available yet. The usecase is a host on the internet whose node exporter can be reached via http-over-TLS, protected by basic authentication.

I tried to hack the existing template, but I can't import my modified version due to some errors whose root cause is unknown to me (I know nothing about Zabbix or its templates): "Import failed: Invalid parameter "/1/host": invalid host name."

Thank you.

Kind regards,

Ralf