June 20, 2014 at 9:02 am
Hello!
We have implemented a Nagios server for monitoring all services of the Company.
In us SQL Server 2008 we have some values in warning and critical status.
One of them is Latches waits x sec.
How I can define the correct value for this? In the sql server i have a database of approximately 300 GB.
I supouse this values change about the type, size, etc of database.
Thanks for the help!
Regards from Uruguay.
June 20, 2014 at 1:32 pm
The only 'normal' that counter has is 'normal for your system'. It's dependant on hardware, database (structure and size), workload and probably other factors.
Monitor for a while, see what is normal for your system, set your alert values to a number above that.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
June 20, 2014 at 2:17 pm
GilaMonster (6/20/2014)
The only 'normal' that counter has is 'normal for your system'. It's dependant on hardware, database (structure and size), workload and probably other factors.Monitor for a while, see what is normal for your system, set your alert values to a number above that.
Ok, thanks for your help.
Then, I do not have a way for determinate the correct value? For example using disk statics, cpu used, etc
Thanks!
Regards.
June 21, 2014 at 6:34 am
There's no such thing as a correct value for that counter. There's normal for your system, and abnormal based on your baselines.
Monitor for a while, see what is normal for your system, set your alert values to a number above that.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
June 23, 2014 at 6:35 am
Ok, thanks.
Another question: if I note the average of latches goes up, what actions I need to take to resolve it? (if it is really a problem as indicated nagios).
Regards.
June 23, 2014 at 6:40 am
Start by investigating what changed.
Also bear in mind that a single counter is almost never useful by itself. If the latch waits/sec changes, what else has changed, in what way. You need to look at the whole picture to troubleshoot something, not a single value.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
Viewing 6 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply