one thing i've always liked to have in my rolodex is someone who can figure out really difficult ops problems, like odd server behavior or database performance issues and take whatever steps are needed.
I used to work with a contractor who knew linux inside and out, used very low level tools to watch packets, OS behavior, etc, and could just about always solve vague questions like "our db is occasionally taking 1 second to pick up a connection, and our dba has been looking at it for a week with no results". Sadly, he got a (great) job and stopped contracting.
If you can accomplish that sort of thing, its worth a high billing rate.
That said, when i've needed that kind of service, the last thing i would ever do is put up a post on stack overflow or something that says "for some reason, our cluster in europe is hozed and we're missing client SLAs. the US one works fine. we already tried rebuilding it. etc etc
Thats because if the problem is that hard, we're not going to realistically get help from the public. we probably won't even be able to provide enough detail to make it solvable via post. and its potentially embarrassing.
I think you'd have to do your some selling to get into people's backup plan.
Maybe put together a portfolio of detailed stories of doing this sort of thing? The more technical detail the more credible, at least for me as a buyer. maybe optimize for searches like "intermittent timeout" :)
maybe a physical object that you mail to VPs of engineering that they stick on their desk until they need it? "Break glass in case of unfixable server issue"