Contextualize Me -- The Case for Context in Reinforcement Learning