[Nagiosplug-help] random DNS check

Philipp Huebner philipp.huebner at credativ.de
Wed Feb 13 08:09:11 CET 2008


Hey,
thanks for your answer.

Thomas Guyot-Sionnest schrieb:
>> Hi list,
>>
>> I want to monitor the performance of our local DNS-Server.
>> If I use check_dns it always checks the same domain which will be cached by our server.
>>     
>
> If you check authoritative servers it should be fairly representative of
> the performance of your DNS servers. If you'd rather like to have a
> global view of what latency your users get around the world (which it
> the only "better" monitoring I could think of for an authoritative DNS
> server) you'll need a distributed service that usually comes as a paid
> subscription (I know Gomez though I'm not sure if they can check just
> the DNS, you can at the very least see the resolving time of an HTTP
> query; maybe Keynote, and there's certainly others...).
>   
I think I didn't give enough details for my situation:
We've set up an internal dns running bind. This dns resolves requests 
(e.g. www.google.com) against the dns of our provider.
It happened some weeks ago, that people hat problems resolving domains, 
it took very long. When they've once reached the website they were 
looking for all was fine, it was the domainname-resolving that has been 
so slow.

My aim is to check the latency of the dns-lookups. But if I check always 
the same domain, our local dns caches it and the time will never be the 
one it takes for our  providers dns to sent the result.
>> So I'm looking for a plugin that works similar to check_dns but checks valid random domains.
>> Maybe I could develop an own plugin for this case, but I don't have the time to do that.
>>     
>
> - From that sentence I'm understanding you're checking a recursive DNS
> server (DNS cache). That's doesn't make much sense, especially if you
> try looking for random domains! The time it will take will be dependent
> of tiers networks and you'll be alerted for THEIR problem. The only
> checks I'd recommend on such servers would be either your own domains
> (you can set-up a special zone/entry with ultra-low TTL if you want to
> force re-check) or one of the root servers (like "a.ROOT-SERVERS.NET.").
> What you're really interested in is knowing that your DNS cache
> responds. You can't really help if the internet is broken...
>   
I try to explain more detailed: to avoid the problem with our local dns 
caching the results, I want to check another domain each time.
Nagios is graphing the results via nagios-pnp, so in the end I'd have a 
graph showing the dns-performance users inside our lan experience.

So this plugin should either use a long list of valid domains and check 
another one each time, or generate several ip's, lookup the domainnames 
recursively, and if this has been successfull, use this domainnames for 
the actual check.

Hope this makes the situation more understandable for you.

Regards,
Philipp Hübner




More information about the Help mailing list