0

I have set up munin to monitor EC2 instances. Now my CPU usage graph (using the unmodified cpu plugin) looks like this:

CPU Usage Graph

The problem is that the "iowait" parameter is returned correctly (see text below the graph), but it is never plotted. On other non-EC2 machines I have never seen this problem. Also there are no error shown in the munin log files.

How can I fix this?

j0nes
  • 955
  • 1
  • 12
  • 27

1 Answers1

1

The problem is that your numbers add up to more than 200, while your graph is max'd at 200. (Have you recently added CPU to this instance?)

Telnet to the munin-node port (4949) and look at the output of config cpu, possibly the result of --upper-limit on the third line: 

graph_title CPU usage
graph_order system user nice idle iowait irq softirq
graph_args --base 1000 -r --lower-limit 0 --upper-limit 400
[...]

If there's something in there that's limiting the graph to only 200%, look in the munin config files (usually /etc/munin-node/plugin-config or the like) to see if it's accidentally hardcoded.

David Mackintosh
  • 14,293
  • 7
  • 49
  • 78
  • David, this is correct. When manually setting upper-limit to a higher value, iowait is correctly shown. One question remains though: I am using the stock plugin that should set its upper-limit to $NCPU * 100. My instance reports 2 CPUs in /proc/stat which is correct. Why does the instance then report data for more than 200% usage? – j0nes Mar 12 '12 at 14:18
  • I'm going to guess that you have multiple cores or hyper-threads or something which will return CPU usage values, but which are not counted as CPUs for some reason. I don't have any EC experience which would explyan why this would happen. – David Mackintosh Mar 12 '12 at 15:09
  • David, this problem seems to be EC2 related: https://forums.aws.amazon.com/thread.jspa?threadID=89271&tstart=0 Accepted your solution because I could fix my munin graph, however I will ask another question regarding how to find out the correct EC2 cpu usage. – j0nes Mar 14 '12 at 07:16