2

Please don't down vote this question as well. I have a genuine question and probably am not phrasing the question correctly. I asked the following question in serverfault and it was closed as "Not a real question".

https://serverfault.com/questions/490317/what-metrics-should-i-monitor-to-guarantee-scalability-high-availability?noredirect=1#comment547719_490317

Would you please let me know where I am offending the moderators? Is shortlisting metrics that indicate the scalability and HA of the website very broad question? Are requests for guidelines out of scope of Serverfault?

Regards, Raj

1 Answer 1

3

MrDenny was the only moderator involved in that decision. The others are just normal users. I saw the question and didn't think it was worthy of a mod-hammer, but he disagreed apparently.

There's really no way to phrase a question like that to make it acceptable for Server Fault. SF is for System Administrator (et al) only, this is a developer question to start with. Metrics that indicate scalability are not necessarily related to availability either. But both are a function of design decisions.

We used to commonly get questions from people (especially those developing the next Facebook killer, or GMail killer, or whatever large website killer) asking roughly the same grade of question. The short answer for all of them is that you should concentrate on making a functional web server first, then work on scalability.

That is not to say that building in a degree of scalability and availability, in the initial design, will not help in the long term. But far too many people become obsessed with making something perfect and lose sight of making anything at all.

As someone who is struggling figuring out where to measure the scalability or availability prospects of a particular design, you should probably simply concentrate on building something functional, and correcting the design as pain-point become obvious.

2
  • Thanks. Actually I already have a website in production that is handling around 1 MM records in 10 minutes (number of concurrent users is low but the volume of transaction is high). The volume is going to increase as we rollout the service for more countries and more products. My specific goal is to start monitoring the website in the production so that I can understand how the site is holding up during peak usage and during new country / product rollout as well as to start collecting metrics and use it to scale the site in preparation for new country / product rollout. Commented Mar 23, 2013 at 7:38
  • 4
    Sounds like you probably need to hire someone who specialises in scalability.
    – dunxd
    Commented Mar 25, 2013 at 11:02

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .