[gsoc-dev] [ganeti_webmgr] "Ganetiviz" Cluster Visualization app added - Ready for pull into "develop" (#5)
pranjal mittal
mittal.pranjal at gmail.com
Thu Sep 5 11:11:44 UTC 2013
On Thu, Sep 5, 2013 at 4:40 PM, pranjal mittal <mittal.pranjal at gmail.com>wrote:
> Hello Chance,
>
> Ah, I get the point! I will prioritize working on a fix for these speed
> issues so that there are no problems in production environments with a lot
> of nodes and vms (data).
>
> Firstly I need to know whether the *speed bottleneck* really lies in the
> fetching the data from GWM or in the rendering of the graph (Case A or B
> mentioned ahead)
>
> *Case A* - Fetching JSON (which indirectly hits the database)
> *Case B* - Javascript code responsible for rendering.
>
> Checking Case A - Does it take some time when you access:
> /ganetiviz/(cluster-slug)/vms?
>
> Checking case B - Does it take a time significantly different than twice
> of Case A to generate a
>
(* significantly *greater* than)
> cluster map when you hit - */map/(cluster-slug)?*
> *
> *
> I would need your feedback after this code is test run on production again.
> *
> *
>
> I could create a test_feature/13569 branch which adds profiling and timing
> to the views so that this is easier to spot the "slow code" location this
> way?
> OTOH, I have spotted areas where ganetiviz Javascript needs to be
> optmized, working on that too.
>
>
>
>
> - Pranjal
>
>
>
>
>
>
> ---------- Forwarded message ----------
> From: Chance Zibolski <notifications at github.com>
> Date: Wed, Sep 4, 2013 at 11:54 PM
> Subject: Re: [ganeti_webmgr] "Ganetiviz" Cluster Visualization app added -
> Ready for pull into "develop" (#5)
> To: osuosl/ganeti_webmgr <ganeti_webmgr at noreply.github.com>
> Cc: Pranjal Mittal <mittal.pranjal at gmail.com>
>
>
> It reduces the number of queries to the database, because each Ajax
> request
> is hitting a view which does a DB query. Since all of this is pretty
> simple
> in terms of data, a single JSON object can easily be created from a single
> DB query. Mostly speed, because with a larger number of nodes and vms, it
> can end up taking some time to load. We've seen this on our staging server
> with production data.
>
> If you want to make the JSON more modular, you can change your views from
> DetailViews.
>
>
>
> On Wed, Sep 4, 2013 at 9:04 AM, Pranjal Mittal <notifications at github.com>wrote:
>
>
> > I was thinking (on the contrary) of making JSON more modular in future -
> > Example: an endpoint /ganetiviz//vms listing the vms for the node. This
> > should help in getting data as and when required rather than fetching a
> > whole sea of JSON data even if its not needed.
> > However, I do not see this ^ being immediately useful, since ganetiviz
> > (currently) needs all that data at once; so what you suggest could be
> > useful, but I am curioous - What is the advantage of returning only 1
> set
> > of JSON?
> >
> > —
> > Reply to this email directly or view it on GitHub<
> https://github.com/osuosl/ganeti_webmgr/pull/5#issuecomment-23801663>
> > .
> >
>
> —
> Reply to this email directly or view it on GitHub<https://github.com/osuosl/ganeti_webmgr/pull/5#issuecomment-23812141>
> .
>
>
>
> --
> Best Regards,
> Pranjal Mittal
> B.Tech. 2014
> Indian Institute of Technology,BHU
> Varanasi, U.P,
> India
>
> Github <http://github.com/pramttl> | LinkedIn<http://in.linkedin.com/pub/pranjal-mittal/26/660/318/>|
> Blog <http://pranjalmittal.in>
>
--
Best Regards,
Pranjal Mittal
B.Tech. 2014
Indian Institute of Technology,BHU
Varanasi, U.P,
India
Github <http://github.com/pramttl> |
LinkedIn<http://in.linkedin.com/pub/pranjal-mittal/26/660/318/>|
Blog <http://pranjalmittal.in>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osuosl.org/pipermail/gsoc-dev/attachments/20130905/a0553571/attachment.html>
More information about the gsoc-dev
mailing list