Question

In my web application I would like to make a “response-test” process. This process can be launched from a web page to see if the Database server is reachable and the approximative time to get some results.

I don’t want precise time of the execution of the stored procedure, I would like to have an idea of the “reactivity” of the Database..

So I make a method like this :

        var start = DateTime.Now;
        var test = new ObjectManager().ListObjects(SomeParameters);            
        var stop = DateTime.Now;
        TimeSpan diff = stop - start;
        label.Text = diff.TotalSeconds.ToString();

In facts the ObjectManager.listObject() is a method who get some datas from database throught a stored procedure.
So the time is calculated for all the process.
Contacting the database, launch stored procedure, cast and retrieve the datas, ..
Not only for the stored procedure. This is not a problem.

This works fine, but I have two problems.

1. I need to make a stored procedure “independent” from quantities of
   datas.

2. I noticed that at the first execution the time returned is for
   exemple 1,62123 seconds. And if I refresh the page the time is
   0,89897 seconds.. Half time! after that each refresh is around 0,8
   secs..

It seems it is caused by sql cache..

So, I need a stored procedure ‘independent’ from quantities of data’s in the Database and which don’t use the sql cache.. Any idea to do that?

Answer 1

You can create a procedure which contains some SQL code that returns a static value. Like SELECT 1; or SELECT getdate()

That way you check the connection to your database without hitting any data.

The first time you run the procedure the SQL server will need to make a plan for it, that will cause some overhead.

Of course you can always run the procedure several times in a row (from code) and calculate an average.