This is really dependent on your department culture.
Officially (at least for all the tenure systems I've participated in), tenure decisions are based on the quality, reputation, and impact of the candidate's research (and to a lesser extent, teaching and service), and their potential for future intellectual leadership and international stature. (Woo, that's a mouthful.) In particular, the decision is based on what the candidate has actually accomplished.
Are they publishing high-quality work in high-visibility venues? Is their work highly regarded in the research community (as described in the letters)? Do they have a solid track record of external funding? Do they have a solid track record of successful graduate student advising? Have they accumulated other indicators of respect within their research community: awards, plenary talks, invitations to program committees/editorial boards, well-placed and successful students? Do they have other evidence of research impact: open-source software, community-standard data sets, or patents?
If this case is strong on these merits, then the day-to-day work habits of the candidate shouldn't matter, and usually don't. In particular, the reference letters, which are by far the most important component of the case, will be written by people who are probably completely unaware that the candidate disappears to a fishing lodge in the Adirondacks every summer. There have been multiple successful tenure cases in my department where the candidate essentially vanished every summer.
More strongly: If their research case is strong, then what assistant professors do during the months the university is not paying them is really none of the promotion committee's business. I work in a field where a three-month vacation in the Adirondacks (with a sufficient supply of pens and paper and reliable wifi) would be amazing for my research productivity.
Nevertheless, faculty are apes; we do apey things. Every department and every field has its own cultural norms, and assistant professors must be aware of and sensitive to those norms. Some departments have a shared belief—which may even be correct in their field—that the research required for a successful tenure case requires being physically present and visibly productive 24 hours a day, seven days a week, 52 weeks a year. Others won't even notice when you're in the office being productive, because they all keep their office doors closed, or they work at home, or in coffee shops, or in a shack in in the Adirondacks. Others have a de facto list of accomplishments that are necessary, sufficient, or both. (NSF CAREER? NIH R1? Three papers in Science? H-index above 50? IEEE Fellow? Letters from five Nobel laureates?)
Whether or not tenure decisions should depend on these local expectations is utterly immaterial. They do. Either you meet them, or you find a job somewhere else. And the only way to understand what these expectations are in your department is to ask directly, and to insist on regular mentorship and feedback from senior faculty.