0

I have a Munin installation which surveils a bunch of servers. Two of them have MySQL replications running of which Munin plots the slave lag by use of mysql_slave_status.

[mysql-slave.nmk.loc]
   address mysql-slave.nmk.loc
   use_node_name yes
   mysql_slave_status.warning 0:2
   mysql_slave_status.warning 0:1800

Only 2 seconds warning level is a test value, because upto 200 seconds are completely normal for this slave - but it does not matter here.

The warning threshold is also visible on the graph plotted by Munin (constant line on height 2). Nevertheless, although the value is much exceeded, * there is no e-mail generated * the frame of the plot remains gray instead of yellow * the status link in the overview remains normal blue instead of yellow

It get even more weird that all other wanings work perfectly. The warnings are generated as mail due to this entry in munin.conf:

contact.emerge.command mail -s "Munin ${var:host}" emerge@monitor.nmk.loc

Can someone help me get Munin warn MySQL slave lags, too?

rexkogitans
  • 324
  • 1
  • 3
  • 22

1 Answers1

0

Self-answer after a couple of hours trial-and-failure and to share knowledge: mysql_slave_status is just the name of the plugin, seconds_behind_master the name of the value; fully qualified bound together, and it works:

[mysql-slave.nmk.loc]
   address mysql-slave.nmk.loc
   use_node_name yes
   mysql_slave_status.seconds_behind_master.warning 0:600
   mysql_slave_status.seconds_behind_master.critical 0:1800
rexkogitans
  • 324
  • 1
  • 3
  • 22