Ad Widget

Collapse

Wishlist and suggestions to zabbix 1.1

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • PerroVerd
    Junior Member
    • Dec 2004
    • 11

    #1

    Wishlist and suggestions to zabbix 1.1

    Hi

    Just a few suggestions very useful for me.

    a) All the graphs are predefined to one hour but you can change dinamically this behaviour, but if you exit the changes are lost. It's possible to save this changed period.

    b) The same in screens

    c) Actions depending on the time or the day of the alert. Ex: If alert1 is activated between 0AM and 12PM send a sms to john if is activated between 12PM and 8PM send a sms to peter and if is between 7PM and 2AM send a mail to support.
    Actually, I can handle this with a catchall mail and processing this mail later, but it would be nice to be included on zabbix.

    d) A tree view of the depend triggers or something similar, i dont know how. I think the dependencies seem confusing.

    e) I have several user-defined parameters (a mix of mon, nagios, mrtg and handmade plugins) . I think its a good idea to have a place to share all, maybe a wiki.

    Well, i thinks that's all

    Thanks in advance
  • limo
    Senior Member
    • Dec 2004
    • 192

    #2
    Voting for WiKi

    I vote for WiKi with user parameters. Could be very usefull..

    Comment

    • joe
      Junior Member
      • Nov 2004
      • 27

      #3
      I modified the PHP to check 2 variables held in include/defines.inc.php.

      define("REFRESH",300); (5 minute refresh)
      define("PERIOD",86400); (24 hour period)

      Most of the pages check those 2 variables to determine how often to refresh, and what the default period should be.

      The patch is based on 1.1alpha4 (and also works on 1.1alpha5).

      period-refresh.diff

      NOTE: I will not guarantee this works for you. It works for me.

      To patch, this is what I did.

      1. Copy the patch to /path/www/zabbix (wherever your zabbix is installed)
      2. cd /patch/www (ie, zabbix parent directory)
      3. Type: patch -p0 < period-refresh.diff

      There may differing methods for this. PLEASE CREATE A BACKUP COPY BEFORE MAKING THESE CHANGES.

      The following files will be modified: charts.php, graph.php, screenedit.php, screens.php, and include/defines.inc.php.

      The period and refresh default to the Zabbix original defaults (1 hour and 30 seconds).

      joe
      Last edited by joe; 28-01-2005, 20:11.

      Comment

      • petkovsc
        Junior Member
        • Feb 2005
        • 6

        #4
        more feature requests

        - It would be nice to have another view into alarms/actions. What actions do
        I have currently set across all hosts?

        - Global settings for alarms:
        - Select a bunch of triggers, set them all to email a group and apply.
        - Search for a particular trigger across all hosts (server is down) and set
        an alarm for all of them.
        - Or even global alarms. When any server {HOSTNAME} is unreachable,
        send an alarm.

        - Alarm after X triggers. I'm using icmp ping to test my vpn availability.
        However icmp drops occasionally. I'm alerted when it goes down and when
        it goes up. It's silly for me to get these two alerts back to back. That
        happened twice this morning. If I could say, alert if it fails 3 times in a
        row and/or 5 times in 10 minutes. Maybe also have an alarm dependent
        upon two triggers, both have to fail before I'll alarm.

        Comment

        • joe
          Junior Member
          • Nov 2004
          • 27

          #5
          Originally posted by petkovsc
          - It would be nice to have another view into alarms/actions. What actions do
          I have currently set across all hosts?

          - Global settings for alarms:
          - Select a bunch of triggers, set them all to email a group and apply.
          - Search for a particular trigger across all hosts (server is down) and set
          an alarm for all of them.
          - Or even global alarms. When any server {HOSTNAME} is unreachable,
          send an alarm.

          - Alarm after X triggers. I'm using icmp ping to test my vpn availability.
          However icmp drops occasionally. I'm alerted when it goes down and when
          it goes up. It's silly for me to get these two alerts back to back. That
          happened twice this morning. If I could say, alert if it fails 3 times in a
          row and/or 5 times in 10 minutes. Maybe also have an alarm dependent
          upon two triggers, both have to fail before I'll alarm.

          This can be done using the max/min values parameter.

          {server:icmppingsec.max(180)}=0

          This will keep the trigger from becoming true until 3 minutes (180 seconds) have passed (6 ping checks with the default 30 second timer) and ALL ICMP was dropped. If even 1 of the 6 attempts worked, then it would not trigger. This is what I have settled on lately.

          I do not think there is a way to check the last X values. You can do last(0) to check only the last value. last(X) would be useful!

          joe

          Comment

          Working...