mibuso.com

Hey,

we have a job which has a long runtime and we want to solve this problem. So I used the client monitor to analyse the job and the sql-queries. I have only access to the client monitor and not to the advanced forms of the client monitor. My question is, which tipps and tricks do you have for me to get a good output from the client monitor. I searched for operations, which have a long "elapsed time". How can I find things, which can be improved? Is it for example a problem, if the entries of the key, filter or sql-index have different orders? I already searched the forum but I didn't found some general tipps.


Best regards!

If you are using MS SQL, you can use SQL Profiler to get info. If there is too much reads for expected count of record, than there is no good index. BUT! Be sure first, that you have maintanance in place, like updating statistics and defragmenting indexes. If there is no maintenance, the results are biased...

Hey,

unfortunately I have no access to the SQL Profiler yet. Therefore I want to use the client monitor. We already defragmented the indices, but the performance of the job is very bad. Also it generates a lot of locks, so I want to use the client monitor. Which other tipps can you give me for the client monitor? Is it for example better / necessary, if / that the key and the sql-index have the same values in the same order?


Best regards!

Look for duration, read count, if it was scan or seek. If scan, look for used index and used filters. There is no one simple answer for performance tuning. You can see it from other posts about SQL performance on this forum...

OK, thanks for your help and tipps. I try to get a little bit deeper in this stuff.

Best regards!