Ad Widget

Collapse

Any volume < 10% free space?

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • obrienmd
    Junior Member
    • Jul 2010
    • 26

    #1

    Any volume < 10% free space?

    Is there a way to configure a trigger for Windows and/or Linux to report if any mounted (non-optical ideally) file system has less than 10% free space?
  • sarathyme
    Member
    • Mar 2009
    • 58

    #2
    You can use the item that retrives % of free space.

    Add an item to monitor free disk percent

    Key -vfs.fs.size[/var,pfree]

    Type - Numeric (float)

    Unit - %

    Set Update interval (e.g. 5 minutes)

    Create a trigger with expression like this
    vfs.fs.size[/var,pfree].max(900)}<10

    This account for free disk space is less than 10% for last 15 minutes. You can change it to last(0)<10 if want to alert immediately.


    Regards
    Vijay

    Comment

    • obrienmd
      Junior Member
      • Jul 2010
      • 26

      #3
      Thanks much, Vijay! Do you know of any way to watch _all mounted_ filesystems for this?

      Comment

      • henry
        Junior Member
        • Jun 2010
        • 16

        #4
        Guess you have to set a trigger for each fs separately.

        h.

        Comment

        • obrienmd
          Junior Member
          • Jul 2010
          • 26

          #5
          I hope there may be a way around this, each of our many servers has different file system setups. Some of our Windows servers have C: and D:, some have C:, and our *nix servers are all over the place with different drives / mount patterns.

          Comment

          • Alexei
            Founder, CEO
            Zabbix Certified Trainer
            Zabbix Certified SpecialistZabbix Certified Professional
            • Sep 2004
            • 5654

            #6
            Originally posted by obrienmd
            I hope there may be a way around this, each of our many servers has different file system setups. Some of our Windows servers have C: and D:, some have C:, and our *nix servers are all over the place with different drives / mount patterns.
            That's in our roadmap for Zabbix 2.0.
            Alexei Vladishev
            Creator of Zabbix, Product manager
            New York | Tokyo | Riga
            My Twitter

            Comment

            • obrienmd
              Junior Member
              • Jul 2010
              • 26

              #7
              Ah, wonderful! I'd be happy to test once it's in trunk!

              Comment

              • henry
                Junior Member
                • Jun 2010
                • 16

                #8
                That's good news, Alexei!

                It'll help me, too, as we have a similar problem with all different file systems and mount points on different Solaris machines.

                Is the reoadmap somewhere on the web, so we can see what else is planned?

                Cheers,
                h.

                Comment

                • jzondervan
                  Junior Member
                  • Jul 2010
                  • 19

                  #9
                  I hope there may be a way around this, each of our many servers has different file system setups. Some of our Windows servers have C: and D:, some have C:....
                  If you add both C and D volumes, the D volume will be disabled if it doesn't exist. (ver 1.8.1)

                  Don't know of a good way to do this on a linux fs, however..

                  Comment

                  • Murilex
                    Senior Member
                    • Nov 2009
                    • 124

                    #10
                    For me, the best way to address this issue was discussed here:


                    We could create a generic template, like:
                    Template Name: Generic_Filesystem_Template
                    Macro (Template): {$PARTITION} => "/"
                    Item Description: Free space on {$PARTITION} (%)
                    Item Key: vfs.fs.size[{$PARTITION},pfree]
                    Trigger Name: Low free space on {$PARTITION}
                    Trigger Condition: {Generic_Filesystem_Template:vfs.fs.size[{$PARTITION},pfree].last(0)}<10
                    Graph Name: Disk utilization on {$PARTITION}

                    Then we could create more specific templates (associated with the Generic_Filesystem_Template), like this:
                    Template Name: Home_Partition_Template
                    Macro (Template): {$PARTITION} => "/home"

                    Template Name: Var_Partition_Template
                    Macro (Template): {$PARTITION} => "/var"

                    After then, we would associate the necessary templates (depending on the partitions existing on the host):

                    Host A (partitions "/" and "/home"): Associated with Generic_Filesystem_Template and Home_Partition_Template

                    Host B (partitions "/" and "/var"): Associated with Generic_Filesystem_Template and Var_Partition_Template

                    Logically, low level macros would take precedence (host macros first, XXX_Partition_Template macros second, Generic_Filesystem_Template macros third).

                    To finish, it would be very good if application were configured globally (the templates above would be configured on the same application, i.e. Filesystem) and all this macros were substituted by their corresponding values on the whole interface, logically considering the precedence discussed above.

                    Comment

                    • QwErTy_LoGiC
                      Member
                      • Feb 2010
                      • 66

                      #11
                      On the roadmap? Sweet!

                      I actually had to create templates for my Windows servers, one for each drive letters...

                      ...Yeah, if you take a peek at our servers, each letter is used at least once...

                      And just to inflict some more pain, we have servers that run in english and some in french... So that two sets of templates (at least for performance counters) for the same drive...

                      ... And to add insult to injury, Microsoft decided to change the wording of their counters in french in 2008... So that's a template in French 2003 and another in French 2008...

                      ...And just to piss me off, I recently discovered that I did not need the exact wording of counters such as \Logical Drive(C\Avg. Disk Queue Length. I could instead use a universal numbering scheme such as \236(C\200 for current disk queue length

                      ...So having a catchall for all disks will be really cool! I'm kinda sick of creating disk templates! But hey, don't get me wrong, template ARE way much better than nothing!

                      Cheers!

                      Comment

                      • Murilex
                        Senior Member
                        • Nov 2009
                        • 124

                        #12
                        I have my doubts about "universal numbering scheme" of windows counters. Recently, I've found two Windows 2003 Servers (R2 - English) with different numbers for terminal services objects representation:

                        Textual representation:
                        Active Sessions: perf_counter["\Terminal Services\Active Sessions"]
                        Inactive Sessions: perf_counter["\Terminal Services\Inactive Sessions"]

                        Number representation (Server A):
                        Active Sessions: perf_counter["\2102\2106"]
                        Inactive Sessions: perf_counter["\2102\2108"]

                        Number representation (Server B):
                        Active Sessions: perf_counter["\3100\3104"]
                        Inactive Sessions: perf_counter["\3100\3106"]

                        I've solved my problem using textual representation, as I have only English Servers.

                        Comment

                        • QwErTy_LoGiC
                          Member
                          • Feb 2010
                          • 66

                          #13
                          Ah dang it...

                          I had my doubts of the reliability of the numbers... I just had not encountered a situation where they did not match...

                          What you're telling me does not surprise me but I was wishing that the issue would not creep up... At least I kept all my textual templates!

                          Thanks for the heads up!

                          Comment

                          Working...