The European XFEL will generate extremely short and intense X-ray laser pulses of high coherence and nearly diffraction-limited divergence. Guiding these X-rays beams over a distance of more than 1 km to the experiments requires an extreme precision in pointing stability of the optical beamline components like mirrors and gratings and also a good control of the divergence of the beam. The specifications of the X-ray mirrors that will be used to transport, distribute and focus the beam are high demanding. It will be required for the reflecting surfaces to have a surface quality of better than 2 nm Peak-To-Valley over a 950-mm length: the ratio between these two parameters, on the order of 10⁻⁹, makes the requirements very challenging to be accomplished. In order to account for the real shape of the mirrors and to assist the production with absolute metrology, it is proposed to use a Fizeau interferometer. Being the mirrors much bigger than the interferometer clear aperture, it is however needed to use an angled (“grazing incidence”) cavity setup to be able to measure the mirrors over their entire length. In using this setup, there are some open questions about the reproducibility of the method, the influence of the particular grazing angle that is used and the level of accuracy that could be expected with different averages. We present a discussion about theory and practical implementation of “grazing incidence” interferometric measurements, with some examples of real measurements at European XFEL on the first beam distribution mirrors.