Actually, I'm not concerned about down time do to FB having a problem.
I'm worried about down time do to the server FB runs on having a
problem. Right now, it is difficult to do a hot fail-over with FB. This
is an issue at large enterprises. Even if FB could run without a
software glitch for years yet could not be easily setup to do a hot
failover, that would be enough to keep it out of some enterprises. This
is the primary reason many databases support clustering; so that a
mission critical application can continue uninterrupted even when "a"
server goes down -- it's not usually to compensate for the server
"software" being unreliable. JMTC.