Re: [Ganglia-developers] Avg Utilization

2008-12-10 Thread Witham, Timothy D
>I believe the graphs should be template driven, somewhat like the
>template pages.

I agree with Jesse that this is rather difficult with current code.

>- whether or not to show stats in the legend

In my graph.php I have (like in trunk):

# Make small graphs (host list) cleaner by removing the too-big
# legend: it is displayed above on larger cluster summary graphs.
if ($size == "small" and ! isset($subtitle))
$rrdtool_graph['extras'] = "-g";

>- whether or not to strip the domain name from hostnames (currently a
>global setting in conf.php)

In my graph.php I have (like in trunk):

# automatically strip domainname from small graphs where it won't fit
if ($size == "small") {
$strip_domainname = true;
# Let load coloring work for little reports in the host list.
if (! isset($subtitle) and $load_color)
$rrdtool_graph['color'] = "BACK#'$load_color'";
}

>There are probably other things that could be configured too.

With current code, we can easily add some value like the above without 
requiring users to hack on templates or config files.

>One related issue is that the HTML pages should specify the height and
>width in the IMG tags for graphs - maybe a function is needed to
>generate IMG tags.

In older ganglia, the HTML set the size to values not equal to the image 
generated.  This forced the images to an unnatural size causing fonts to miss 
pixels and other ugly effects.  So sizes were removed to let the browser use 
the actual image size.

-twitham

--
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers


Re: [Ganglia-developers] Avg Utilization

2008-12-10 Thread Witham, Timothy D
> IMO the graphs are already to cluttered for a quick overview and I have
>ideas for adding even more clutter (that matters to me, like the date the

Good idea to add the end date/time to graphs larger than default (like when you 
zoom an individual graph to save it).  I think we can do this in graph.php.  It 
is not needed on default graphs since you see the time on the HTML page and 
don't need this duplicated on all the graphs.

>graph was produced :-). But if clutter doesn't matter, why not do both? I

Where do you draw the line?  I want to know average value of all metrics 
plotted, not just CPU or load.  Why is memory less important?

If we put all the numbers in text on the left, then they have to be labeled, 
duplicating the labels already in the graph legends.  Since the graphs already 
have color-coded legend labels, it seems logical to add the mean values right 
there where the user is already looking.  In some cases, this adds one text 
line (14 pixels?) to the image.  In other cases it adds none.

Either way, the fact that I can instantly see the average value of all the 
colored squiggly lines and areas, instead of having to eyeball a guess against 
the Y axis, makes up for the 14 pixels lost in the plotting area.  We could 
even add 14 pixels to default graphs if that is a concern.  It would be harder 
to go looking for the corresponding number in the text to the left of all the 
graphs.

>do not believe that the additional overhead is such a problem.

True.  But it seems there has been some questions as to whether the 
calculations are being done correctly, and where the data comes from (load or 
CPU).  That's why I think it is cleaner to leave only the current load data in 
the text, and add all average numbers to the graph images where the source is 
obvious.

-twitham


--
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers


Re: [Ganglia-developers] Avg Utilization

2008-12-10 Thread Jesse Becker
On Wed, Dec 10, 2008 at 10:22,  <[EMAIL PROTECTED]> wrote:
>> This is slightly circular, since rrdtool will take height and
>> width arguments for the size of the chart--*not* including
>> the border, title, legend, etc--and return the total image
>> size on the command line.  However, since we use passthru(),
>> this information is currently lost.
>
> That suggests that rrdtool needs some extra command line option telling
> it that the dimensions are for the image size.

Newer versions of rrdtool have this.  I meant to mention that previously.

-- 
Jesse Becker
GPG Fingerprint -- BD00 7AA4 4483 AFCC 82D0  2720 0083 0931 9A2B 06A2

--
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers


Re: [Ganglia-developers] Avg Utilization

2008-12-10 Thread Jesse Becker
On Wed, Dec 10, 2008 at 03:51,  <[EMAIL PROTECTED]> wrote:
>>  IMO the graphs are already to cluttered for a quick overview
>> and I have ideas for adding even more clutter (that matters
>> to me, like the date the graph was produced :-). But if
>> clutter doesn't matter, why not do both? I do not believe
>> that the additional overhead is such a problem.
>
> I believe the graphs should be template driven, somewhat like the
> template pages.

While this would be nice, it would be very using the current web front
end.  The HTML template system currently in use is basically a
"one-way" deal:  the various PHP scripts process various things (.rrd
files, taling to gmetad, etc), and stuff the results into pre-existing
HTML template files.  The templates themselves do not do anything
besides presentation.  Specifically, nothing in the templates are used
as "input" for these calculations.

> However, the choice of templates would replace the choice of graph sizes
> - so the user would be able to choose which template they want.
>
> Each template could specify:
> - graph size

This is only possible because of clever (ab)use of HTML forms and URL
parameter processing.  The other ideas you mentioned (show stats,
fonts, hostnames) cannot be driven explicitly at the template level,
given the current code.

> There are probably other things that could be configured too.

Of course! :)

> One related issue is that the HTML pages should specify the height and
> width in the IMG tags for graphs - maybe a function is needed to
> generate IMG tags.

This is slightly circular, since rrdtool will take height and width
arguments for the size of the chart--*not* including the border,
title, legend, etc--and return the total image size on the command
line.  However, since we use passthru(), this information is currently
lost.

A further complication is that crafting the various reports is not
always easy.  Trying to make them flexible enough to be truly template
driven will make the code much more complicated.

Now, all of that said, I actually don't object to the idea of making
the web frontend more flexible.  In fact, I think that it is a great
idea.  But there's very much a tradeoff between making/keeping the
code sane, and doing everything that we'd like.


I propose something of a middle ground:  make the various parts of the
frontend generic enough that building a new template is straight
forward.  Specifically, make graph creating simpler, and in such a way
that it is easy to specify a number of these different
parameters--sizes, titles, etc.  If things are set in the conf.php
file, perhaps we could let a different .php file override some of the
settings on a per-template basis?

-- 
Jesse Becker
GPG Fingerprint -- BD00 7AA4 4483 AFCC 82D0  2720 0083 0931 9A2B 06A2

--
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers


Re: [Ganglia-developers] Avg Utilization

2008-12-10 Thread Martin Knoblauch


- Original Message 
> From: "Witham, Timothy D" <[EMAIL PROTECTED]>
> To: Brad Nicholes <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]" <[EMAIL 
> PROTECTED]>; "ganglia-developers@lists.sourceforge.net" 
> 
> Sent: Monday, December 8, 2008 5:48:31 PM
> Subject: Re: [Ganglia-developers] Avg Utilization
> 
> >> Windows users are looking at the figure and thinking that `Avg
> >> Utilization' refers to CPU utilization (from the cpu_report graph).
> >>
> >> Maybe both are needed:
> >>
> >> cluster_util_load: displayed as `Avg Utilization (Load)'
> >>
> >> cluster_util_cpu: displayed as `Avg Utilization (CPU)'
> >>
> >> Can anyone suggest a better way to name these figures, or would this be
> >> an acceptable patch?
> 
> I would remove all that and put the numbers on the graphs only, like done in 
> trunk.  I think it is clearer to have the % numbers on all of the graphs.  
> That 
> way the user sees average values for all metrics plotted, right there in the 
> graph legend.  So they can look at CPU or Load or Memory or anything, instead 
> of 
> wondering what the number off to the side is.
> 
> I have voted this way in the 3.1 STATUS file.
> 

 can't vote at the moment, but I really like the display of the average(s) on 
the overview pages. My vote is "+1".

 IMO the graphs are already to cluttered for a quick overview and I have ideas 
for adding even more clutter (that matters to me, like the date the graph was 
produced :-). But if clutter doesn't matter, why not do both? I do not believe 
that the additional overhead is such a problem.

Cheers
Martin


--
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers


Re: [Ganglia-developers] Avg Utilization

2008-12-08 Thread Witham, Timothy D
>> Windows users are looking at the figure and thinking that `Avg
>> Utilization' refers to CPU utilization (from the cpu_report graph).
>>
>> Maybe both are needed:
>>
>> cluster_util_load: displayed as `Avg Utilization (Load)'
>>
>> cluster_util_cpu: displayed as `Avg Utilization (CPU)'
>>
>> Can anyone suggest a better way to name these figures, or would this be
>> an acceptable patch?

I would remove all that and put the numbers on the graphs only, like done in 
trunk.  I think it is clearer to have the % numbers on all of the graphs.  That 
way the user sees average values for all metrics plotted, right there in the 
graph legend.  So they can look at CPU or Load or Memory or anything, instead 
of wondering what the number off to the side is.

I have voted this way in the 3.1 STATUS file.

-twitham


--
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers


Re: [Ganglia-developers] Avg Utilization

2008-12-08 Thread Brad Nicholes
>>> On 12/8/2008 at 5:38 AM, in message
<[EMAIL PROTECTED] 
, <[EMAIL PROTECTED]> wrote:

> 
> I've been getting a few questions about the Avg Utilization displayed on
> the web pages.
> 
> Searching the code, I found it defined like this:
> 
> meta_view.php:  $cluster_util = sprintf("%.0f", ((double)
> find_avg($clusname, "", "load_one") / $avg_cpu_num ) * 100);
> meta_view.php:if (isset($cluster_util))
> meta_view.php:   $tpl->assign("cluster_util", " size=\"-1\">Avg Utilization (last $range):"
> meta_view.php:
> ."  $cluster_util%");
> 
> In fact, the figure represents the average of load average values for
> the period graphed.
> 
> Windows users are looking at the figure and thinking that `Avg
> Utilization' refers to CPU utilization (from the cpu_report graph).
> 
> Maybe both are needed:
> 
> cluster_util_load: displayed as `Avg Utilization (Load)'
> 
> cluster_util_cpu: displayed as `Avg Utilization (CPU)'
> 
> Can anyone suggest a better way to name these figures, or would this be
> an acceptable patch?
> 

This sound appropriate to me.  

Brad


--
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers