I am sure this has been asked multiple times - but I could not find a good answer anywhere. Hence this question.
I have a azure cloud service that reads data from a SQL Server 2012 Database. The primary use of this data base is Read-Only. The data gets populated once every month and the cloud service only reads from it (SELECT) (Yes - I could move this to a key-value store - but as of now I cannot change the way things are)
What would be a good way to estimate the number of SELECT operations that this database can handle per second. The machine on which this runs is a 4 core, 7GB ram azure vm. I am just looking for some ballpark numbers to reasonably estimate the number of users my service can handle. The table is clustered on the column which I will filter (WHERE) results on.
Would greatly appreciate help with this.
Specifying more details based on a reply below:
What level of performance is acceptable?
My end-user SLA should be around 2 seconds. So I expect the DB to return in approx 100 - 500 milliseconds
How many rows are being returned from these Selects?
- On an average I might be returning around 50 rows per select call. Total data being < 200KB
Size of the database
- Around 15M records totalling around 150GB
I understand that this is a vague question - would be great to get some numbers from folks based on their experiences.