Upper limits and confidence intervals are a convenient way to present experimental results. With modern experiments producing more and more data, it is often necessary to reduce the volume of the results. A common approach is to take a maximum over a set of upper limits, which yields an upper limit valid for the entire set. This, however, can be very inefficient. In this paper we introduce functional upper limits and confidence intervals that allow to summarize results much more efficiently. An application to upper limits in all-sky continuous gravitational wave searches is worked out, with a method of deriving upper limits using linear programming.