#680 new defect

Can't override warning/critical from munin.conf for negative fields

Reported by: jxh Owned by: nobody
Priority: normal Milestone: Munin 2.2
Component: master Version: 1.2.5
Severity: normal Keywords:


Using SNMP to graph if_err_, the plugin has two fields, recv and send, with default warnings as follows:

recv.warn -12500000
send.warn 12500000

I want to override these both to zero in munin.conf so I see if any errors at all are detected, e.g.:

    use_node_name no
    snmp_rgi1_if_err_2.recv.warning 0:0
    snmp_rgi1_if_err_2.send.warning 0:0

This works for send, but not recv, apparently because recv is "hidden" under the "send.negative recv" configuration in the plugin. (I happen to have two hosts getting errors, one only on send and the other only on receive, and only one of them is yellow.)

Change History (5)

comment:1 Changed at 2009-10-21T02:06:10+02:00 by janl

  • Milestone set to Munin 1.4

comment:2 Changed at 2009-11-15T00:03:35+01:00 by janl

  • Milestone changed from Munin 1.4 to Munin 1.4.1

comment:3 Changed at 2010-03-08T13:56:42+01:00 by janl

  • Milestone changed from Munin 1.4.4 to Munin 1.4.5

comment:4 Changed at 2012-01-24T16:35:49+01:00 by snide

  • Milestone changed from Munin 1.4.7 to Munin 2.1

comment:5 Changed at 2013-04-17T16:28:23+02:00 by somenet

This still is an issue!

Recently a switch behind a switch died therefore the monitored interfaces didnt go down, but because of the vlan configuration there was no incomming traffic on eth1 anymore, leading to some legacy (=windows, without any monitoring) systems becomming unreachable.

I have hoped to get a critical warning, because i had set up the interfaces so that if no traffic in any direction passes, there must be something broken.

But because clients were still trying to reach the subnet, the up never went to 0, just there was no response and down = 0 for like 8 hours until someone call'd.

    if_eth1.down.critical 1:
    if_eth1.up.critical 1:

I played with the plugin a little, and actually managed to get the down display as critical.

if_eth1.down.graph yes

solved the problem and the critical show'd up, but what also show'd up: the down traffic in the graph as negative and now positive line too.

Therefore i think one solution would be to introduce a new value for xxx.graph called hidden, meaning:

  • xxx.graph yes : show line description at graph-html-page + show line in graph. (as is now)
  • xxx.graph no : show line description at graph-html-page + hide line in graph.
  • xxx.graph hidden : hide line description at graph-html-page + show line in graph + hide critical/warning information. (as is xxx.graph no atm)
Munin version: 2.0.12
Last edited at 2013-04-17T16:31:26+02:00 by somenet (previous) (diff)
Note: See TracTickets for help on using tickets.