Sometimes there are known issues with some datasets which cannot be fixed but users should be aware of them with high priority to avoid related problems.
lfpw (MF)
sea ice (ci) is available only up to day +60 in contrary to +61 for other params
ammc (BoM)
wrong maximum values of surface_air_maximum_temperature_sfc at some points
@Frederic Vitart: The completely wrong max temperature from Australia in af ew coastal| grid points is a know problem. I had an exchange of emails with the Australian about a year ago. There is no quick fix since it is present in the re-forecasts and all the re-forecasts would need to be rerun. A possible solution was to mention in the Australian model documentation that tmax should be used only over land points (using ammc land-sea mask) since the problem happens only over sea points.