The solar UV (UltraViolet) flux, especially the EUV (Extreme UltraViolet) and FUV (Far UltraViolet) components, is one of the main energetic inputs for planetary upper atmospheres. It drives various processes such as ionization, or dissociation which give rise to upper atmospheric emissions, especially in the UV and visible. These emissions are one of the main ways to investigate the upper atmospheres of planets. However, the uncertainties in the flux measurement or modeling can lead to biased estimates of fundamental atmospheric parameters, such as concentrations or temperatures in the atmospheres. We explore the various problems that can be identified regarding the uncertainties in solar/stellar UV flux by considering three examples. The worst case appears when the solar reflection component is dominant in the recorded spectrum as is seen for outer solar system measurements from HST (Hubble Space Telescope). We also show that the estimation of some particular line parameters (intensity and shape), especially Lyman a, is crucial, and that both total intensity and line profile are useful. In the case of exoplanets, the problem is quite critical since the UV flux of their parent stars is often very poorly known.