[44064] in North American Network Operators' Group
Re: Reportable Metrics
daemon@ATHENA.MIT.EDU (Iljitsch van Beijnum)
Fri Nov 2 12:26:03 2001
Date: Fri, 2 Nov 2001 18:26:48 +0100 (CET)
From: Iljitsch van Beijnum <iljitsch@muada.com>
To: "Murphy, Brennan" <Brennan_Murphy@NAI.com>
Cc: "'nanog@merit.edu'" <nanog@merit.edu>
In-Reply-To: <B481990C9658D411BD3C009027D6F544024772EF@SNC-5-87.nai.com>
Message-ID: <20011102182021.J75854-100000@sequoia.muada.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Errors-To: owner-nanog-outgoing@merit.edu
On Fri, 2 Nov 2001, Murphy, Brennan wrote:
> What am I missing?
If you're going through all this trouble, you might as well measure some
interface statistics such as:
- CRC errors, these are an important clue indicating lower layer problems
- collisions, if you use any non-switched ethernet (or even if it's
switched: duplex mismatch is a bad thing and it happens)
- over- and underruns, these indicate (transient) high CPU loads
- input/output drops, to see if you are experiencing congestion
And:
- router CPU load
These should all be easy to measure with MRTG if you can find out how to
read the info from the box using SNMP.
