When saying how good someone WAS in a particular year, ERA is what I'll use. Why wouldn't you? It's not to say there isn't cause for concern moving forward, but I'm not necessarily all that concerned, although if he's coveted, I wouldn't be opposed to moving him.
What I mean is that runs are the most important when evaluating a reliever's effectiveness for a given year, but certainly not for projecting their future. I will say, there are obviously other factors that would be points of contention when ranking relievers, so you couldn't just go by ERA. Things like stranded runners, high/low leverage, etc.
Like you say, you also look at stranded runners and leverage. In addition, I'd look at unearned runs since a pitcher is generally partially partially responsible for them.
But besides all of that, I would also look at the peripherals and stats such as FIP, xFIP, etc. I think those numbers do a better, or at least as good of a job of telling us how a pitcher actually did compared to ERA. ERA strongly factors in defense and luck. I would also consider strength of defense, the parks played in, and the strength of competition.