Surveys (more commonly known as "reviews") give an overall introduction to the field. They don't necessarily contain all the work that has been done in history, if it is an old field, but they should contain all relevant recent work and present the state of the art. See e.g. this source or this source.
It's pretty common for review articles to be invited papers, where the journal's editorial board asks some famous researcher to write the review. But you can definitely write your own. The peer review process is similar to standard research articles. You submit it, the journal reviews it, and returns a decision. One cannot really talk about "chances of acceptance" - that's completely dependent on how good the paper is.