Wednesday, January 18, 2006

getting results? how do you know?

your friendly neighborhood happyfunboy received some pretty awesome news today.

had my 3 month assessment since starting to exercise regularly (3 times a week) last october.

since that time, yours truly has:
  • lost 5 pounds
  • increased lean mass (muscle) by 5 pounds
  • decreased fat mass (fat, obviously) by 5 pounds
  • reduced % body fat from almost 25% to less than 20%
  • lost 2.5 inches at the waist
all the way around, solid progress. be sure, i may have felt this progress in other ways, or noticed anecdotal evidence of becoming more my pants constantly falling off me anymore.

but the difference is...

i not only know for sure, i also know exactly how much progress i've made.

and how do i know all this?

that's simple...

before i started working out, i opted to take advantage of the free assessment offered to new members.

and as part of that assessment, one of their exercise physiologists took certain measurements of me...measurements which now serve as my starting point.

and it got me thinking about something almost no one does in sbs-land. of hands here at the funcave...

who among the folks here actually baselines their servers?

i'm betting that number is pretty low, especially if your shop isn't running separate app servers or database servers.

heavy-duty performance tuning really requires baselining.

otherwise you risk making some change that not only could fail to increase performance, but could actually bring a system to its knees.

well, you know what?

the same goes for all the not-so-heavy-duty performance tuning you prolly should be doing.

if someone reports sluggish response, are you capturing enough data to really know what is going on with that system?

or are you just blindly fumbling in the dark, hoping to hit the switch that turns on the lights...and not the one that sets off the fire alarm?

to be honest, it doesn't even have to be a lot of info.

for a server, the minimum items to track to get an idea of performance are:
  • disk
  • memory
  • processor
  • network/lan
but beyond just have to know what to look for, by determining what values should be considered out of the norm.

and just what is the norm?

well, it can be...and usually is...different for every environment, based on server size, configuration, load, etc. etc. etc.

so why do this at all?

it's not just the drastic decreases in performance that can kill you.

in fact, because those kinds of issues are normally noticed and addressed right away, i'd argue they are even less costly.

it's the little bit here, little bit there performance hits that can one day set you back on your heels, wondering holy christmas! what fracking happened?

and that's where the baseline matters so much.

otherwise, you might never notice your progression, or regression, as the case may be.

and even if you wouldn't be sure.

which is the difference between being a professional...or a wannabe.


Blogger Vlad said...

I don't know that its fair to compare it between a professional and a wannabe as a wannabe in this context means someone that would want to install servers.

Nothing stops completely inept people from deploying servers. Or trying to. How many disasters have you been called in on, with server down, no passwords, no docs, etc.

At the end of the day it matters little what Chris thinks of you. It mattes even less what I think of you. But if you cut corners and don't know what you're installing, don't establish baselines, don't document your deployments, don't practice disaster recovery, don't plan for capacity and LOB... well, what we are saying here hurts a lot less than that weekend you'll have to spend in an empty office fixing a server because you didn't spend the extra five minutes up front to do it right.

Congrats on the weight loss, I am sure many of the SBS Show fans from Cell Block H will appreciate the good news.

8:15 PM  

Post a Comment

Links to this post:

Create a Link

<< Home