Announcement

Collapse
No announcement yet.

Graphs with holes

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Graphs with holes

    We use in our structure CENTREON combined with Nagios to monitor 400 services today. We plan to increase this number with goal to monitor a maximum of services about 1000 at the end. To implement new service ce made the choice to add step by step.

    All worked fine until we reached the number of 485 services. Till we have 500 services monitored holes appeared in the graphs. We tried to increase the number of services checked in nagios.cfg and we also tried to increase the number of lines in the ndo buffer configuration. Nothing works.

    Has anyone an idea of the parameter to check or how to diagnose the problem ?

    Any help would be appreciated.
    Attached Files

  • #2
    Hi,

    1. What are the server specifications ? Is it a VM ?
    2. Do you use the NDO Centreon patch ? This patch reduce the nb of entries in the censtatus database, so reduce the nb of request.

    I already had this problem with FAN (Fully automated Nagios). When i moved to a manual install of nagios / centreon and applied the patch, the problem disappear.

    Look at the documentation on how to patch NDO http://en.doc.centreon.com/Setup:ndoutils

    Cheers
    Ubuntu server 10.04 LTS 64 Bits - Haute dispo 4 serveurs centraux (réplication MySQL + VIP + Rsync)
    Nagios 3.3.1 | Centreon 2.3.9 | Centreon-Broker 2.1.1 | 2000 hôtes - 5000 services | 6 Remote Pollers

    Dev : CES 2.2 - Centengine - Centreon 2.4.1

    Comment


    • #3
      Originally posted by magicbret View Post
      Hi,

      1. What are the server specifications ? Is it a VM ?
      2. Do you use the NDO Centreon patch ? This patch reduce the nb of entries in the censtatus database, so reduce the nb of request.

      I already had this problem with FAN (Fully automated Nagios). When i moved to a manual install of nagios / centreon and applied the patch, the problem disappear.

      Look at the documentation on how to patch NDO http://en.doc.centreon.com/Setup:ndoutils

      Cheers
      Hi,

      Thank you for your fast response

      1) Server spécifications

      No VM
      Intel Xeon E5530 2,4 Ghz
      RedHat 5.6 64 bits
      Ram : 2 Go

      2) I have done a manual installation of Nagios and centreon

      I have installed the patched version of ndo : ndoutils-1.4.b9 with patch ndotuils-1.4b9_light.patch

      3) It is the only solution to reduce requests ?

      4) A beginner question : how do you add specifications at the bottom of text ?

      In Advance thank you

      Best regards

      Comment


      • #4
        Hi,

        Install NDO tools module (http://forum.centreon.com/showthread...O-Tools-Module) and try the low execution time query inside + the last post query's (not added in module atm).

        Have you monitor your server disk I/O? if no, do it. (http://exchange.nagios.org/directory...rmance/details)
        Haven't you any very verbose debug enabled?

        What about your MySql status (slow query...)?

        are your holes with regular freq? random? in relation with a high rate of errors?
        Raphael
        --
        Bi Intel(R) Xeon(TM) CPU 3.06GHz - 3Gb ram
        Debian
        Nagios® Core™ Version 3.2.1 - Nagios Plugins 1.4.14 - NDO 1.4b9 patché
        Centreon 2.3.4 - Syslog Module 1.3.2 - StatusMap Module 2.5 - NDO Tools Module 0.4 SVN - Nagvis
        Beta tester : centreon-engine - centreon-broker

        Comment


        • #5
          Originally posted by raphaelt View Post
          Hi,

          Install NDO tools module (http://forum.centreon.com/showthread...O-Tools-Module) and try the low execution time query inside + the last post query's (not added in module atm).

          Have you monitor your server disk I/O? if no, do it. (http://exchange.nagios.org/directory...rmance/details)
          Haven't you any very verbose debug enabled?

          What about your MySql status (slow query...)?

          are your holes with regular freq? random? in relation with a high rate of errors?
          Hi

          My CENTREOn version was 2.0.2

          I tried to install the ndo module. Installation was succesfull but no data was displayed.

          I have finaly upgraded CENTREON plaform to 2.2.2 version. Graphs are continus now with the same number of services monitored 485. The ndomodul displays data. Could you give me precisions about : try the low execution time query inside + the last post query's.

          I haven't a verbose debug on.

          Attached as image the load of the server. I will now implement the check_cpu to be more fine. Will publish it when data got.

          I can't figure out were was the problem and when I will encounter it one more time (if ?)

          The holes appeared randomly.

          No particular errors seen in logs.

          Is there a rule in number of parallel checks that I can launch ? How to determine the max number ?
          Attached Files

          Comment


          • #6
            latency on centreon is generaly about CPU, Load, RAM and/or I/O.

            high check latency is increase load
            very verbose mode or big database increase I/O
            ...

            i never heard about max parallel check number because it s fonction of your hardware (when i was on a hold P3 server if i increase upper than 400 services perf was falling down. now i have more than 1000 services on my bi-xeon)
            Raphael
            --
            Bi Intel(R) Xeon(TM) CPU 3.06GHz - 3Gb ram
            Debian
            Nagios® Core™ Version 3.2.1 - Nagios Plugins 1.4.14 - NDO 1.4b9 patché
            Centreon 2.3.4 - Syslog Module 1.3.2 - StatusMap Module 2.5 - NDO Tools Module 0.4 SVN - Nagvis
            Beta tester : centreon-engine - centreon-broker

            Comment

            Working...
            X