Chief ray angle is the angle the main ray likes to hit the pixel, especially in the corner of the chip. Microlenses on the sensor are aimed. If the lens sends its corner ray at a different angle, the well sees less light, more colour shading, and a softer edge. People call this "the corners went magenta" and then blame the ISP.
In one sentence. The sensor has a preferred incoming angle. The lens has an outgoing one. If they disagree, the centre of the picture is a liar.
Why small pixels are fussier
A large old pixel would accept a sloppy ray and still fill. A 1.4 µm BSI pixel with aggressive microlenses will not. Wide M12 glass often has a high CRA at the edge. That is the usual fight: a wide, short lens on a small, modern chip.
The skylight test
Stand under a skylight and look up. In the middle the sky is bright. If you walk to a deep window reveal, the same sky is a dim slit. The light did not change. The angle did. A pixel with a tight microlens is that window reveal.
How to pair it
Read the sensor CRA if the datasheet publishes it. Then open the lens library and look at Maximum CRA. They should be in the same neighbourhood, not "both exist." If either side is silent, treat corner colour and corner brightness as a measurement you still owe yourself, not as a surprise in PPAP.