The SQL Server Deadlock-Collector collects information which can be read out from the ring buffer in the memory of the SQL Server, and stores this information in a database.

Aloof from the pure deadlock graph, the data is parsed, and the first 2 respective processes of the deadlock graph are contrasted in 2 columns, respectively, with information such as “process ID, SQL/procedure text, host name, application name” etc.

This way, one obtains a tabular representation of the deadlock processes. Since there may be more than 2 involved processes, the complete graph is always also present.

With the aid of a few sample reports, administrators will obtain a quick overview on the frequency of the occurring deadlocks in the respective databases and be able to pass on the data to the developers.

Developers will thus obtain a valuable and quickly interpretable version of the occurring deadlock combination and be able to implement prevention strategies with this information.

 

Deadlock Graph

WaitResource_LockMode_IsolationLevel

QueryPlan

Report_Top20InvolvedResources

Last edited Feb 27, 2015 at 1:30 PM by AndreasWolter, version 5