Covariance change in COV matrix , does it really make the change the way i think?

https://lh5.googleusercontent.com/vPsKM-iVgEjhrsqLTcjIjcbv2lK6y3z-FHl0KUzjTL86Vukq-NWFMIuew-qyT18sagrT8y7wTk9taokikVcC=w1920-h937

I understood the covariance concept as the change in one variable with respect to other. If we think about sigma(xy) as positive ,then x axis should tilt with a slope of 0.8 , if i am right?

Then the variation of points should be around that y axis and the x axis. Please clarify more about multivariate normal distribution and covariance concept as i want to understand it with some practical usage of that.