Haven't we? vs Don't we?
Short answer: 1 is common in American English; 2 is common in British English; 3 is uncommon in British and American English; 4 is common in British and American English.
Tag questions (American English), or question tags (British English) are often taught as simple rules, but their use is actually rather complicated.
The simple rules are:
The English tag question is made up of an auxiliary verb and a pronoun. The auxiliary must agree with the tense, aspect and modality of the verb in the preceding sentence. If the verb is in the present perfect, for example, the tag question uses has or have; if the verb is in a present progressive form, the tag is formed with am, are, is; if the verb is in a tense which does not normally use an auxiliary, like the present simple, the auxiliary is taken from the emphatic do form; and if the sentence has a modal auxiliary, this is echoed in the tag:
- He has read this book, hasn't he?
- He read this book, didn't he?
- He's reading this book, isn't he?
- He reads a lot of books, doesn't he?
- He'll read this book, won't he?
- He should read this book, shouldn't he?
- He can read this book, can't he?
- He'd read this book, wouldn't he?
- He'd read this book, hadn't he?
https://en.wikipedia.org/wiki/Tag_question
According to those rules, because "we have plenty of time" is in the present simple, the correct tag would be do: "we have plenty of time, don't we?" This is the tag that's typically used in American English.
However, the common (a common?) rendering of the sentence in British English uses have as the tag: "We have plenty of time, haven't we?"
Gunnel Tottie and Sebastian Hoffman have written a very interesting study of the differences between American and British tag questions: "Tag Questions in British and American English".
Here's their summary, with added emphasis on a particularly relevant finding:
Summary and Discussion
We have shown that the use of “canonical” tag questions differs in spoken colloquial British and American English in some spectacular ways:
First of all, there are nine times as many tag questions in British English as in similar types of American English (but the frequency of tags is lower in less colloquial British English).
The preferences as regards polarity types differ in the two varieties, the greatest difference being that negative–positive tag questions are more frequent in American English than in British English.
The choice of auxiliaries or modal verbs in tags also differs, with American English preferring DO-tags and British English HAVE-tags; this can be explained by the differences in the use of tenses in the two varieties and the predilection for HAVE GOT constructions in anchors in British English.
Both dialects show a great variety of verb–pronoun combinations, 200 different types in all. The claim that invariant innit is taking over thus seems premature.
Our results concerning the pragmatic functions of tag questions must be regarded as preliminary, but there appear to be substantial differences between the two varieties here as well. Facilitating tag questions account for a greater proportion in American English, and confirmatory and attitudinal uses account for a greater proportion in the British data. Aggressive tag questions are used only by British speakers, but in a very low proportion—only 1 percent of all cases. It is probable that their saliency explains the attention that this numerically small category has received in the literature. We have not been able to correlate polarity types and pragmatic functions, but it is clear from Kimps’s (2005) research that positive–positive tag questions are not predominantly used in aggressive functions, as has previously been claimed.
Speaker age is an important conditioning factor in both varieties, with older speakers using more canonical tag questions and younger speakers significantly fewer. Data from Stenström et. al. (2002) strongly suggest that younger speakers of British English prefer invariant tags like yeah, eh, okay, right, and innit, and we therefore hypothesize that the total use of tag questions and other hedging devices may still be fairly similar across age groups.