On 2017-06-12 16:14, Tim Stiller wrote:
Best performance Monitoring Tool ever: http://nabox.tynsoe.org
That is good, yes. But the key Q is always: what is it that someone intends to do with the [latency] numbers once they are there and can be seen? If you have no idea what the numbers mean, how to interpret them, or indeed if they are good or bad, then what is the point of measuring/tracking s**tloads of counters? There are 1000s of them inside ONTAP.
I know this sounds like I'm just generally challenging the usefulness of any and all performance metrics tracking and in a way I am -- in another way I'm not. *I* want to see those numbers and try to learn from them, but I generally do not want any non-experts to see them, that's my basic attitude
Explanations of causality between performance utilistaion and various latency metrics in a highly complex technical system for the purpose of budgeting for Capacity Management is a daunting task. I feel it has to be kept very very simple, so simple that any layman can grasp it. Pointy Haired boss in the IT industry cannot comprehend complex systemic relations, he's not interested IYKWIM
/M