This includes:

 * moving the SLA upwards, because it's shorter
 * adding an "issues" section to document known problems and a link to
   the bugtracker
 * mention the current service maintainer as an approver
 * remove the launch checklist, that should be in a ticket instead
 * mention benchmarks in the alternatives evaluation

i couldn't figure out what to put in "security", this seems more like
something in the "must have" (duh).

the disaster recovery is part of a larger launch plan and makes no
sense on its own. it also overlaps with the "as built" design docs.

This is the full mounty, and most services will not have everything in
there. But it should cover most of the stuff described in the
checklist (#30881).