[collectd] apache and collectd plugin

[Amit Gupta]

On Tue, Apr 21, 2009 at 6:13 PM, Florian Forster <octo at verplant.org> wrote:
> Hi Amit,
>
> On Tue, Apr 21, 2009 at 02:12:20PM +0530, Amit Gupta wrote:
>> This is what I thought earlier but then I wasn't sure if the
>> scoreboard output (along with legends) will always be present. I will
>> probably re-look at the scoreboard output.
>
> well, currently the detected server software is only *used* within the
> scoreboard code, so if there is no scoreboard present, we don't need the
> detection..

We would have issues in case where say apache is just reading and
lighttpd is reading post, legends for both is  "R". In this particular
case, we wouldn't be able to make determination about the server type.
What should be the type instance in this case?

How about not making any server determination and creating scoreboard
rrd files only if the counter for the legend has a value greater than
zero. As far as conflicts of the legend goes, we will be always
submitting "reading" for the type_instance "R" and lighttpd_reading or
something similar for "r". The consumer of the rrd files will read the
"reading" rrd file for lighttpd as readPost. I agree it is not very
clean but that is the easier solution I can think of :)

Offcourse, if we want to have separate type_instances for both apache
and lighttpd, then there is no way but to determine the server type.

Regards
Amit

>
>> > Oh, one more minor issue: The variable `sending' is submitted with
>> > the instance “sending” on Apache and “write” on Lighttpd. Maybe it'd
>> > be a good idea to unify this..
>>
>> I have kept the type_instance (apache-scoreboard-<type_instance>) as
>> per the legend definition in the mod_status output (which is why I
>> didn't unify it), for instance, the legend 'W' stands for "sending
>> reply"  in apache and "write" in lighttpd. Do you want it to be
>> unified?
>
> I'm unsure.. It'd probably make graphing and comparisons easier, but on
> the other hand it's probably easier to understand the graphs when we
> stick to the terms used by the software queried..
>
> Regards,
> -octo
> --
> Florian octo Forster
> Hacker in training
> GnuPG: 0x91523C3D
> http://verplant.org/
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
>
> iD8DBQFJ7b+MHdggu3Q05IYRAuPAAKCh+b34pON3y6pj/W8etft59ZZaWgCeNpuJ
> tK7Y1jfAyL5JuxnYjToO1Xo=
> =G0hZ
> -----END PGP SIGNATURE-----
>
>