[Nagiosplug-help] Output of check_http not consistent

[Matthias Eble]

Hi Ingo,
>> But it wouldn't get confused if it used perfdata instead of the  
>> plugin output. Thus plugin output is for human beings, perfdata for  
>> computers :)
> good point! So let me paste the whole output:
> HTTP OK HTTP/1.1 200 OK - 1227 bytes in 0.067 seconds | 
> time=0.066948s;;;0.000000 size=1227B;;;0
> OK - HTTP/1.1 302 Found - 0.723 second response time  | 
> time=0.723189s;;;0.000000 size=365B;;;0
> 
> So first the standard map-file *does* get confused, since it is  
> searching for a line starting with "HTTP" - a pattern which does not   
> always fit as we can see.
okay.. but there is afaik no specification about how the output *must* 
look like. The guidelines claim it *should* look like
SERVICE STATUS: Information text
So you are right that it isn't consequent. But to me that's not the main 
problem.
>And second: Are you sure, that the time in  
> both cases is measuring a comparable process/roundtrip? I do not want  
> to compare apples with horses (don't know the English expression for  
> "Birne" right now :-) ).

Beside the fact that the one is on localhost, you should consider to 
monitor the url the redirect actually redirects to (getting status 200) 
because the actual site will likely be more complex to generate (script 
language??) or simply has more content to transfer. I just took a short 
look at the code, but the measured time range should contain the 
get/post request as well as the transfer time of the reply if I saw it 
right.

To summarize: 302 is imho not comparable to 200s because
- 302s are simpler to generate for the webserver
- have nearly no content to be transmitted

hth
Matthias