I have batches of timeouts in SQL and system_health contains around 20 entries all waiting on LCK_M_IX. Majority of entries does not have SQL text entry associated with it ("Unable to retrieve SQL text").
I can run profiler and wait for next occurance (couple of time a day) but this is heavily used server and hence I can not log a lot of information. So which counter with which thresholds I need to put to find out why those locks are acquired and held?
View Complete Post