March 10, 2009 at 10:41 am
I have been asked to create an indicator that would be a blend of SQL Counters which would reflect SQL Server load or stress at given interval (hourly). This indicator would run from 1 (least load) to 10 (highest load). The requestor would be able to look at that number as a general load measure throughout the day
Anyone else ever encounter such a request?
March 10, 2009 at 11:01 am
Set up a range on the various measures. For example, work out what the lightest IO load is (probably 0), and the highest acceptable IO load (will depend on your system). Do the same for CPU, etc. Query the numbers, average the result, you'll have your number.
Given all that, I'm not really sure such a single-score performance monitor is actually worth the effort. Might be, not sure.
- Gus "GSquared", RSVP, OODA, MAP, NMVP, FAQ, SAT, SQL, DNA, RNA, UOI, IOU, AM, PM, AD, BC, BCE, USA, UN, CF, ROFL, LOL, ETC
Property of The Thread
"Nobody knows the age of the human race, but everyone agrees it's old enough to know better." - Anon
March 10, 2009 at 12:11 pm
I'm not certain how meaningful it is either, but sometimes you can tell 'em but you can't tell 'em much.
March 10, 2009 at 12:25 pm
I know how it goes. Good luck on it.
- Gus "GSquared", RSVP, OODA, MAP, NMVP, FAQ, SAT, SQL, DNA, RNA, UOI, IOU, AM, PM, AD, BC, BCE, USA, UN, CF, ROFL, LOL, ETC
Property of The Thread
"Nobody knows the age of the human race, but everyone agrees it's old enough to know better." - Anon
March 10, 2009 at 12:47 pm
I'm not sure how you could do it either, but when you get it done, write it up as an article 'cause I'd love to read it.
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt
Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning
March 10, 2009 at 6:26 pm
Something that queries and records some wait stats (from sys.dm_os_wait_stats) and a couple of counters from master.dbo.sysperfinfo (e.g. page life expectancy) and reports on the hourly differences.
To get a single number you would have to average a few numbers and scale it into the 1 - 10 range.
As has been said I'm not sure how useful this would be, but it sounds interesting nevertheless.
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply