A new study came out in Nature a couple of weeks ago that assesses multiple records of ocean temperatures over the last couple of decades and finds that there is “a statistically significant linear warming trend for 1993-2008 of 0.64 W m-2″.
The challenge the paper took on was one of assessing the uncertainties and inconsistencies in the various records. The paper is Lyman, J.L., et al. 2010. Robust warming of the global upper ocean. Nature 465 and Real Climate had an article about it here.
Also in that issue of Nature is an article by Kevin Trenbreth [PDF] that discusses that paper and sheds some light on his infamous “travesty” comment in the stolen CRU email.
So, oceans are warming on decadal timescales, as expected in an enhanced greenhouse effect climate, and most of the ocean (below 700m) is poorly monitored making short term behaviour in the parts we can see difficult to explain. It is important in understanding global warming to keep the relative heat capacity of the surface and the oceans in perspective.

This graph is from Murphy et al, 2009, h/t to Skeptical Science’s post measuring the earth’s energy imbalance