General rule is the chord of the fairing should be aporoximately 4.25 times the thickness of the tube you want streamlined.
I'm going expound on this statement. The only one I know of that put a "number" to it was Hoerner in one of his aerodynamics books. For struts that number was a c/t ratio of 3.7 or t/c of 27%. That is based on fairing thickness NOT tube diameter or member being faired thickness.

I thouroghly respect Mr. Hoerner and all he has done for aviation. But, some of his 1960's experiments and modelling have been vastly expanded upon by the likes of NASA (as a major example).

It has been found that more elongated shapes are more effective (better laninar flow) - the widest point more in the range of 40-50% from the leading edge (think more like a cigar). And streamlined symetrical airfoils with a length of up to t/c 9% are also more effective (diminishing returns beyond that). 9% probably not very practical and elongated streamlined shapes more difficult to build and heavier, so may not be worth the effort for likely minimal gain. So as a guide, I would say min. 27% to max. 9% t/c, keeping weight to a minimum.

I do agree with Lowell that the fuselage (door) to wing intersection could benefit a lot from streamlining that intersection.