graphite-dev team mailing list archive

[Pete Emerson <question144005@xxxxxxxxxxxxxxxxxxxxx>]

New question #144005 on Graphite:
https://answers.launchpad.net/graphite/+question/144005

Our /opt/graphite/storage path is a RAM disk for performance reasons.

Just recently, our metrics have started to lag behind, and since we use the metrics to do alerting, our alerts are going off.

Here's an example; note that the oldest metric hasn't been updated in 9 minutes:

$ date ; find /opt/graphite/storage/whisper/instances/prod/foo -type f -exec ls -l {} \; | sort | cut -d / -f 1 | uniq -c
Thu Feb  3 22:54:19 UTC 2011
    130 -rw-r--r-- 1 root root 12292 Feb  3 22:46 
    476 -rw-r--r-- 1 root root 12292 Feb  3 22:47 
    781 -rw-r--r-- 1 root root 12292 Feb  3 22:48 
    456 -rw-r--r-- 1 root root 12292 Feb  3 22:49 
    150 -rw-r--r-- 1 root root 12292 Feb  3 22:50 
    492 -rw-r--r-- 1 root root 12292 Feb  3 22:51 
    555 -rw-r--r-- 1 root root 12292 Feb  3 22:52 
    793 -rw-r--r-- 1 root root 12292 Feb  3 22:53 
     28 -rw-r--r-- 1 root root 12292 Feb  3 22:54 

I'm not sure what the bottleneck is; metrics are being sent every minute. Do you have suggestions for settings to tweak or things to look for in the logs that might give me clues?

Pete


You received this question notification because you are a member of
graphite-dev, which is an answer contact for Graphite.