After being bitten by the blocking bug in the past where the index rebuild jobs were blocked by a connection from the application that wasn't closed, I implemented a standard practice to include a lock time out in my script that rebuilds the indexes: SET LOCK_TIMEOUT 900000; -- Set it to 15 minutes to avoid causing blocking

So now instead of the index rebuild job being blocked, it just times out after 15 minutes and moves on to the next index.