1

Background

I was given trace logs from SQL Profiler that was run for 15min during a timeframe where we have been experiencing 100% CPU on the server hosting SQL. I imported those files to a table called TraceTable. I did find some possible causes that are not related to Audit Logout, however, the top-most result is from Audit Logout and I'd like to understand why the reads / writes / CPU are so high for only 15min.

Research

I read that the duration of Audit Logout is some factor involving the culmination of time, not just the time from the trace time frame.

But I can't find anything to explain the high read / write / cpu, other than some general statements that Audit Logout is usually not the cause of performance issues. I also found an answer that says the CPU utilization is cumulative like the duration, but the article linked in the answer doesn't explicitly say what the answer says it says. I would like to verify that this is unrelated to the performance issues we're seeing.

Why does Audit Logout show these concerning stats?

Here is the query I ran:

SELECT TOP 100
  COUNT(*) AS TotalExecutions,     
  EventClass, 
  CAST(TextData as nvarchar(2000)) AS Query ,
  SUM(Duration) AS DurationTotal ,
  SUM(CPU) AS CPUTotal ,
  SUM(Reads) AS ReadsTotal ,
  SUM(Writes) AS WritesTotal
FROM 
  TraceTable
GROUP BY 
  EventClass, 
  CAST(TextData as nvarchar(2000))
ORDER BY 
  CPUTotal DESC

And here is the top result:

result

(Sorry, I don't have access to the SQL instance so I don't know the exact version. It's safe to assume it's at least 2014, possibly newer)

1 Answer 1

2

The Duration, CPU, Reads, etc on Audit: Logout are the totals for the session. This is clearly documented.

So these totals include all of the values from the other events generated over the course of that session.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.