graphite-dev team mailing list archive

[Nicholas Leskiw <question166527@xxxxxxxxxxxxxxxxxxxxx>]

Question #166527 on Graphite changed:
https://answers.launchpad.net/graphite/+question/166527

    Status: Open => Answered

Nicholas Leskiw proposed the following answer:
1.) More details would be useful. How many datasets are you graphing?
How many points per hour are you graphing? How

Have you looked at Disk I/O as the bo

-Nick

On Jul 31, 2011, at 5:15 AM, Steve Thompson
<question166527@xxxxxxxxxxxxxxxxxxxxx> wrote:

> New question #166527 on Graphite:
> https://answers.launchpad.net/graphite/+question/166527
> 
> This is probably more of a Python question than a Graphite question but I thought I'd ask for other users experiences and advice. My Python skills are not far past the "Hello world" stage (although I'm getting better!)
> 
> When I first started playing with Graphite I installed it on a VMWare VM (Ubuntu 10.10). Unsurprisingly, once I started putting lots of data in to it the rendering slowed down. What I found though was that even by adding memcache and giving the VM more processors there was not much of an improvement to the rendering. 
> 
> Now that my company has decided we're going to adopt Graphite I've been doing a few experiments with the hardware I have available. I've taken 3 large datasets I created (a little script that just creates a text file of steve.metric.1 <random number> <timestamp every 5 secs>) and uploaded it to a fresh Graphite install on either Ubuntu 10.10 or RHEL 6 on four bits of hardware: a Intel Core 2 Duo desktop, an AMD Opteron server, an Intel Qaud core server and what we call "Dodecs" - 3 quad core AMD processor machines.
> 
> Although you can see a difference in the processor usage - as in the dodecs clearly aren't working as hard as the Core 2 -  the end result - the time it takes to render various graphs - doesn't seem to change much. 
> 
> My assumption is that something somewhere isn't using the available processing power - for example, Apache is but Python isn't - which is why we observe less processing power for the same end result? But this is an assumption based on very solid facts, the main one being that I don't know the inner workings of Python.
> 
> In summary, do I need to do something to take advantage of the extra processing power or should I not bother putting Graphite on high powered boxes?
> 
> Thanks in advance!
> 
> Steve
> 
> -- 
> You received this question notification because you are a member of
> graphite-dev, which is an answer contact for Graphite.
> 
> _______________________________________________
> Mailing list: https://launchpad.net/~graphite-dev
> Post to     : graphite-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~graphite-dev
> More help   : https://help.launchpad.net/ListHelp

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