GreyShade has responded to the call for the left to develop a clear position on when intervention is justified and why with a typically thoughtful piece examining some of the key models and applying them to Afghanistan and Iraq. One of the models he considers is a consequentialist approach of "justification by outcome" - that a war is justified "whenever the outcome after the war is better than it would have been without the war". He points out that this very quickly runs into problems with countries being judges in their own case, and suggests the UN is a less partial judge (but also a far from perfect one). But there's a bigger problem, shared by consequentialism in general, of what Dan Dennett in Darwin's Dangerous Idea called "computational intractability": in many cases, we don't know what weight to assign to particular outcomes in our calculations of consequences - or even what sign we should give them - not because of uncertainty about possible consequences, but because of uncertainty about how those consequences should be morally interpreted. Dennett illustrates this with the example of Three Mile Island:
How could Three Mile Island have been a good thing? By being the near-catastrophe that sounded the alarm that led us away from paths that would encounter much worse misadventures - Chernobyls, for instance. Surely many people were fervently hoping for just such an event to happen, and might well have taken steps to ensure it, had they been in a position to act. The same moral reasoning that led Jane Fonda to create the film The China Syndrome (a fictional near-catastrophe at a nuclear plant) might lead someone rather differently situated to create Three Mile Island
If we are to use a consequentialist analysis, we need to know at minimum which column to put things in. But even with hindsight we just don't know yet whether Three Mile Island has had good or bad consequences. And foresight is much more difficult.
I think this significantly counts against consequentialism as a war of determining whether a war is justified (and against consequentialism in general). Fortunately, we have other options.