First verify that the rotation is not due to a loose yoke. Check for the existence of electronic rotation adjustments, as LCGuy suggests.
If the yoke is not loose, then the drift in rotation is electrical in nature. A 20-degree rotation is somewhat larger than what most such circuits typically provide, but maybe this monitor is atypical. If that circuit has failed, perhaps one of the adjustment pots has gotten dirty, or the DC supply that feeds it has gone south. If the rotation eventually seems to stop changing, then simply rotating the yoke might be an acceptable fix, but without knowing (and fixing) the true cause, there's always the risk of the drift re-emerging at some point.
If the yoke is not loose, then the drift in rotation is electrical in nature. A 20-degree rotation is somewhat larger than what most such circuits typically provide, but maybe this monitor is atypical. If that circuit has failed, perhaps one of the adjustment pots has gotten dirty, or the DC supply that feeds it has gone south. If the rotation eventually seems to stop changing, then simply rotating the yoke might be an acceptable fix, but without knowing (and fixing) the true cause, there's always the risk of the drift re-emerging at some point.