[email protected]
[Top] [All Lists]

Bug#412826: marked as forwarded (smokeping: graph rendered incorrectly (

Subject: Bug#412826: marked as forwarded smokeping: graph rendered incorrectly (bad scaling?)
From: Debian Bug Tracking System
Date: Wed, 28 Feb 2007 13:30:50 +0000
Your message dated Wed, 28 Feb 2007 15:22:53 +0200
with message-id <[email protected]>
has caused the Debian Bug report #412826,
regarding smokeping: graph rendered incorrectly (bad scaling?)
to be marked as having been forwarded to the upstream software
author(s) [email protected]

(NB: If you are a system administrator and have no idea what I am
talking about this indicates a serious mail system misconfiguration
somewhere.  Please contact me immediately.)

Debian bug tracking system administrator
(administrator, Debian Bugs database)

--- Begin Message ---
Subject: Re: Bug#412826: smokeping: graph rendered incorrectly bad scaling?
From: Niko Tyni
Date: Wed, 28 Feb 2007 15:22:53 +0200
On Wed, Feb 28, 2007 at 03:04:07PM +0200, Niko Tyni wrote:
> On Wed, Feb 28, 2007 at 01:03:45PM +0100, Oskar Liljeblad wrote:
> > Package: smokeping
> > Version: 2.0.9-2
> > Severity: normal
> > 
> > Take a look at the two attached graphs. They are generated from the same 
> > data.
> > Values in the graph from 11:00 to ~13:00 ranges from ~30 ms to ~87ms.
> > If you check the same duration in the second graph, the values now range 
> > from
> > ~30ms to ~235ms.
> > Could it be that in the "zoomed in" graph (10:00..13:00) isn't scaled 
> > correctly,
> > and that the top grey bar that doesn't fit completely is not rendered at 
> > all?
> Hi,
> looking at the code, the upper scale limit is chosen (inside
> Smokeping::findmax()) as 1.5 * the maximum median of the pings with
> the interval. In this case, the maximum median in the zoomed case is
> around 60, so the upper limit is 90. Values that don't fit in this
> apparently get discarded.
> There are the 'unison_tolerance' and 'max_rtt' options that affect this,
> but they only seem to cut the upper scale further down. I assume that
> the 'logarithmic' option will get you all the smoke, but that has some
> obvious drawbacks...
> I'm CCing Tobi Oetiker, who wrote this part of the code. Tobi, is this
> a bug or feature? 
> (The graphs can be found at http://bugs.debian.org/412826 ).

(Tobi: forgot the CC first time, here it goes for real. Please keep
[email protected] in the recipient list.)

--- End Message ---
<Prev in Thread] Current Thread [Next in Thread>