[prev in list] [next in list] [prev in thread] [next in thread] 

List:       rhq-devel
Subject:    discussion on measurement baseline calculation & problem metrics
From:       hrupp () redhat ! com (Heiko W ! Rupp)
Date:       2010-08-03 12:55:01
Message-ID: 9CDB1F6B-204F-484C-83B7-9931394C0359 () redhat ! com
[Download RAW message or body]


Am 02.08.2010 um 21:20 schrieb Joseph Marques:
> 
> [I know that Heiko has had some thoughts about using a more sophisticated algorithm \
> for this sliding window calculation.  I'm sure he'll chime in on this thread.]

As a "real fix" has been already been discussed away, I'll not add anything here...

What I also find missing is the ability to pin a border of the base band.
Take the case of a hard drive getting full or software getting slow. 
This will produce an OOB event. Then the baseband gets re-calculated which means the \
former outlier now defines the border and a "good" state. So while one is informed \
about the OOB once in a while, the situation can slowly get worse without anyone ever \
noticing (the outlier % just needs to be so small, that it shows up on page2 of the \
oob subsystem view).


-- 
Reg. Adresse: Red Hat GmbH, Otto-Hahn-Strasse 20, 85609 Dornach bei M?nchen
Handelsregister: Amtsgericht M?nchen HRB 153243
Geschaeftsf?hrer: Brendan Lane, Charlie Peters, Michael Cunningham, Charles Cachera


[prev in list] [next in list] [prev in thread] [next in thread]