graphite-dev team mailing list archive

[c_t <question219015@xxxxxxxxxxxxxxxxxxxxx>]

Question #219015 on Graphite changed:
https://answers.launchpad.net/graphite/+question/219015

    Status: Answered => Open

c_t is still having a problem:
Thanks for the quick reply!

I know that and I did that. The config has not been changed since the
first install of graphite so no whisper-files were present before.

However I have a rough idea of what could be the actual problem: I
realized that the default interval that diamond uses is 300 seconds.
However I have the retention definition:

retentions = 10s:1d,1m:30d,15m:1y

so the smallest interval here is 10s which means that only every 30th
datapoint will actually have a non-null value. Since I did not customize
the storage-aggregation.conf I guess the problem is the default setting
of "xFilesFactor: 0.5" is used when aggregating 10s metrics to 1m
metrics and this ratio is obviously never reached when diamond sends
data only every 5 minutes.

I have changed the retention schema to

retentions = 30s:1d,3m:30d,15m:1y

and the diamond push interval to 30 seconds which means that now all the
highest-resolution datapoints in graphite should be non-null and
aggregation should work.

I will confirm if that fixed the problem tomorrow after 24 h have
passed. (And yes, I have removed all existing .wsp-files after editing
the retention schema.)

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