Looks like there was a problem in the stored proc being called inside of the cursor, and as you suggested the whole thing being inside of the try catch was...

re: moving the declare / deallocate outside of the try. That might help, thanks. In the mean time I ditched the cursor for another method and it seems to be...

The dataset will never be large. And "slow" is all relative. I know cursors get a bad rap but defending them is not really why I posted...

tweaked the catch block to return the error message (duh), but now i'm even more confused:

There is no other looping going...

Sure, I could use something else other than a cursor. maybe that'd fix the issue. But... in this case the dataset i am looping through is small so a cursor...

Clearly if the only thing that has changed is the size of the column--even if it's a variable length column--that is causing the problem.

Use this query to check the size...

So my coworker figured it out... When you "pause" a job the SQL Agent Job is deleted... It gets recreated when you un-pause the job. Mystery solved!

When the column is bigger, the index is bigger. Bigger indexes take more space on the disk and take longer to create.