We’ll give it some time…

From: Fred Grieco [mailto:fredgrieco@yahoo.com]
Sent: Thursday, October 29, 2015 3:27 PM
To: Rue, Randy; 'toasters@teaparty.net'
Subject: Re: Anybody using OnCommand Performance Manager (OPM)?

If this is opm 6.x, it's using a standard deviation model and not a fixed threshold.   So it's very chatty until new loads have burned in for a reasonable amount of time (a few weeks).   The alerts cam be disabled but I believe that's global. 

---

From:"Rue, Randy" <rrue@fredhutch.org>
Date:Thu, Oct 29, 2015 at 6:22 PM
Subject:Anybody using OnCommand Performance Manager (OPM)?

Our new 8.3 cluster is in production, finished migrating the last vservers to it last weekend, and suddenly we're getting ~6 alerts a day from an OPM server set up by a co-worker who's left the organization. I confess I've barely looked at the thing until now.

Whenever I go to investigate the details of the latest "incident" I find that it's not a problem, i.e. latency on a volume "jumped" from <1ms to ~5ms. Looks like the application is looking for deviation from a dynamically determined baseline and I'm not seeing any option for tweaking parameters or limits.

Six false alarms a day makes this effectively useless.

Anybody else using this app? Got any tips on making it useful?

Randy

_______________________________________________
Toasters mailing list
Toasters@teaparty.net
http://www.teaparty.net/mailman/listinfo/toasters