graphite-dev team mailing list archive

[Naresh V <question160576@xxxxxxxxxxxxxxxxxxxxx>]

Question #160576 on Graphite changed:
https://answers.launchpad.net/graphite/+question/160576

Naresh V proposed the following answer:
On 2 July 2011 02:05, Nicholas Leskiw
<question160576@xxxxxxxxxxxxxxxxxxxxx> wrote:
> Question #160576 on Graphite changed:
> https://answers.launchpad.net/graphite/+question/160576
>
> Nicholas Leskiw proposed the following answer:
> 1.) You're only sending data once a minute, but you have a 10-second
> retention rate.  Any reason for that?

This example I pasted in a 1-minute cron-based metric (echo "metric
val epochtime" | nc graphite 2003)
The other log-tailing metric scripts send whenever there's a pattern
match in the logs (may or may not be every 10s) and they do _not_ face
the same issue (confirmed just now):

% whisper-fetch.py
/opt/graphite/storage/whisper/com/mailhostbox/outbound-us1/postfix-maia/delay/arrival_to_last_active_queue.wsp
 --pretty --from=1309493420 --until=1309493700
Fri Jul  1 04:11:00 2011        0.000000
Fri Jul  1 04:12:00 2011        0.001667
Fri Jul  1 04:13:00 2011        0.020000
Fri Jul  1 04:14:00 2011        0.010000
Fri Jul  1 04:15:00 2011        0.008333

> 2.) Check your message sending script / application.  Send the raw
> messages to a file and to Graphite.  See if your data source is doing
> something strange.

The postgres metrics sending script just does some SELECTs every 1
minute and sends it to graphite. Quite straightforward.

> 3.) Upgrade to trunk.

I've been tracking trunk and I merged from trunk again at the time of
commenting here for the first time -- there wasn't any update to
carbon/whisper though. I'm restarting carbon-cache.py right now in any
case to see if that helps.

Thanks again,

-Naresh V.

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