Internet Connectivity Monitoring

[Giles Coochey]

Hi Alex,

 

Unfortunately most of our upstreams implement CPP (Control Plane
Policing) to prevent DDoS attacks, which essentially means that given a
certain threshold of ICMP messages the routers begin to drop those
packets purposefully to avoid flooding the routers control plane. I do
monitor those, but have notifications turned off because of the false
positives.

 

The other problem is that one of the scenarios I see once in a while is
one of our upstreams may withdraw all their BGP routes and while BGP
converges we may see momentary loss to certain sites. We use BGP
communities to define which provider we use for different geographic
locations, so having geographic monitors is quite important for us
(unfortunately, not important enough to host our own servers in those
geographic locations, nor to use a performance monitoring service such
as Keynote for the same).

 

Even worse, thankfully only happened just once, one of our upstreams
encountered an issue where they lost connectivity, but didn't withdraw
their BGP routes from our tables. (Ouch!).

 

Cheers,

 

Giles

 

________________________________

From: Alex Dehaini [mailto:alexdehaini at gmail.com] 
Sent: 25 January 2008 11:12
To: Giles Coochey; nagios-users at lists.sourceforge.net
Subject: Re: [Nagios-users] Internet Connectivity Monitoring

 

We monitor our upstream provider core routers. Basically, we monitor
every physical device between us and them. So, if any device goes down,
we can tell where the problem is originating from.

 

Secondly, if you have a co-located server outside your network, you can
monitor that as well.

 

Alex

On Jan 25, 2008 8:54 AM, Giles Coochey < gcoochey at sapphire.gi> wrote:

Hi List,

As part of our Internet Connectivity monitoring I monitor a set of 
websites on the Internet from a diverse number of locations
geographically.

There are some issues with this:

(a) I have to avoid sites that use Content Delivery Systems such as
Akamai, because as we host such boxes ourselves the websites actually 
load from local servers which we host rather than the intended
geographic location. Most large sites such as www.yahoo.com and
www.google.com use these content delivery systems.
(b) I have to avoid small sites which often go down due to lack of
redundancy in their implementation.
(c) I have to avoid sites which might get pretty annoyed that I'm 
pulling their homepage once every 10-15 minutes just to see if I'm
getting good access.

I would like to hear from others on the list as to what they do to check
Internet Connectivity and whether they do any smart stuff to include 
multiple connectivity checks as a single service check.

What do you do?

Thanks

Giles

------------------------------------------------------------------------
-
This SF.net email is sponsored by: Microsoft 
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________ 
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users 
::: Please include Nagios version, plugin version (-v) and OS when
reporting any issue.
::: Messages without supporting info will risk being sent to /dev/null




-- 
Alex Dehaini
Developer
Site - www.alexdehaini.com
Email - alexdehaini at gmail.com 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.monitoring-lists.org/archive/users/attachments/20080125/e49b92c3/attachment.html>
-------------- next part --------------
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
-------------- next part --------------
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null