Over the last 100 years or so, globally averaged surface temperature, which is always varying a little, has gone both up and down, but over the whole period it is estimated to have risen about half a degree centigrade (using the US National Climate Data Centre's analysis; other analyses give as much as 0.65C).
However, this value is associated with substantial error bars, and the warming is occurring in a system that can vary about that much without any forcing at all - something not surprising in a system that is both turbulent and heterogeneous.
Yes, there does appear to be warming, but the amount is hardly certain or indisputable. And the amount found does not appear that alarming.
The alarm, I would suppose, comes from the notoriously inadequate climate models.
Ah yes, my favorite chestnut from my years as a graduate student: proof by computer simulation. It was shoddy science then, and its shoddy science now. True science should be based on empirical observation, and models verified by empirical observation (ie models that make measurable predictions that can then be checked against reality to ascertain the model's accuracy). To call global warming alarmism "science" is a gross distortion.