[mrtg] Re: Availability of specific devices

[Rich Adamson]

> Stefan.Heins at seb.de wrote:
> 
> > For the management i have to proof the availability of specific devices.
> > Our contract says that the devices has to be availaby for 98,5%.
> > How can i do this with mrtg? Does somebody out there has the same problem
> > already solved?
> 
> Availability can't be monitored.  You can make a close approximation.
> 
> If you poll the device every five minutes, there is a chance that
> you don't get to see an outage of upto five minutes.

That all depends upon exactly what the contract spells out. If the contract
has been written specific to a device or interface, it can be measured
within some very acceptable limits. If the contract is written in terms of 
some end-to-end language, that's somewhat more difficult.

For example... write a script that polls sysUpTime (1.3.6.1.2.1.1.3) and save
the returned value. The next time polled, if the new value is less than the
old value (excluding 32-bit counter roll-over every 497 days), the box was 
rebooted. The new value is the number of milliseconds since it happened 
(regardless of the polling interval). Calculate the outage and save so mrtg 
can 'accumulate' the values. The shorter the polling interval, the more 
accurate the result will be.  For most equipment, one minute polling should 
be more than adequate for most SLAs. If the box is being rebooted so often 
that this approach can't be used, there is different very serious problem.

If the contract can be reduced to specific device interfaces, use the same
approach but poll ifLastChange (1.3.6.1.2.1.2.2.1.9.x, where x is the specific
snmp interface index needed).

In both cases above, the accumulated 'seconds of outage' can be saved and
plotted via mrtg for that device or interface.  The amount of error is very
small on a monthly SLA, and should have no impact if your objective is
only 98.5% (648 minutes per month).

Or, if there are Cisco routers involved, write a script to use Cisco's 
remote ping facility via snmp (eg, ping address "a" from remote router 
"b" and return the result). If the "a" address is one of your customer's
devices (eg, router, server, switch), and the "b" address is the most
distant router, then the result will take into account any redundant
paths, hot spares, etc. 

--
Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
Archive     http://www.ee.ethz.ch/~slist/mrtg
FAQ         http://faq.mrtg.org    Homepage     http://www.mrtg.org
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi