Clifton,

I think what Dan is asking for is some kind of "predictor" based on 
historical data.  Say 4 weeks ago, utilization was 3%, then 3 weeks 
ago it was 5%, 2 weeks ago it was 8%, and last week it was 10%.  
I know that this is not a perfect world, but I would say with some level 
of certainty that this week will be about 12%, and next week will most 
likely be in the neighborhood of 15%.  It would be neat to show a 
straight line that would show this "trend" over the next "x" weeks or 
until the trend shows a certain percent of utilization (or maybe an 
"SLA average exceeded" predictor?).  I'm think about how to do this 
one now... 

Paul

>>> Clifton Royston <[EMAIL PROTECTED]> 03/21/01 02:32PM >>>
I think you need to define a little better exactly what you want this
trend line to graph.  I didn't quite get it from this description.  It
might turn out that it is doable already, or is something somebody
could tell you how to do if you clarify a little more.

"Growth towards the service level" means what?  Maximum value ever seen
to that point in time (maximum of maximum)?  Exponential rolling
average of the maximum?  Or did this tie in to the percentile
discussion because you want it to graph the 95th percentile over a
rolling time window?

  -- Clifton

-- 
 Clifton Royston  --  LavaNet Systems Architect --  [EMAIL PROTECTED] 
   WWJD?   "JWRTFM!" - Scott Dorsey (kludge)   "JWG" - Eddie Aikau


--
Unsubscribe mailto:[EMAIL PROTECTED]
Help        mailto:[EMAIL PROTECTED]
Archive     http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi

Reply via email to