March 7, 2017 at 7:49 am
I've been reading a lot about this and haven't been able to find a good definitive way to measure how much of its max capacity a server is currently using, all the numbers I see about transactions per second, memory,cpu, are not really accurate or do not apply to all the scenarios, i.e and OLTP box can't be measured in the same way as an server for reports or analytic.
I would like to hear some ideas
Thanks in advance
March 7, 2017 at 8:28 am
I think you may have the cart before the horse here... For most companies, the definition of "maximum capacity" might well differ. Some will want to be sure that a given system NEVER runs above some percentage of the "absolute maximum". Others are looking to live rather closer to the edge, and squeeze every last ounce of capacity from a server that they can. Without acknowledging where you are in that regard, even the meaning of maximum capacity might not do much for you. Once you establish that, the meaning of maximum capacity becomes a bit more clear, and may then provide the basis for useful measurements. After all, it doesn't really provide much value to say that a server is operating at 80% of maximum capacity this month, and at 75% next month, and at 85% the following, without some kind of context to go along with it. Were there variations in business volume that can account for the differences? How about infrastructure issues, such as network problems? Sometimes, I even wonder if it might be better to measure "useful maximum capacity", as often, if you actually use 90 to 95% of actual maximum capacity, response time can suffer significantly. So basically, when you boil it down, you need to take a lot more into account than just some absolute figure representing maximum capacity, as often, actually using all of it can be more trouble than it's worth.
I think it's more useful to measure things that add value, such as availability, or response time, or some form of throughput. These measures can usually be directly tied to business activity in some fashion, whereas the term "maximum capacity" means different things to different people, and doesn't necessarily do much for you that the specs on the server hardware couldn't do. That said, however, if you put the same stuff on two different but identically spec'ed servers and one screams while the other flails, then maximum capacity may not be the only thing in play, but it would at least establish a basis for comparison.
Steve (aka sgmunson) 🙂 🙂 🙂
Rent Servers for Income (picks and shovels strategy)
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply