We’ve written elsewhere about more general availability and good systems design, but what if you want to assure availability when things go unexpectedly wrong?

Read more The Register