Here is a sampling of MS theses.
It is a good idea to have models on which to base your thesis. All theses are measured both by thoroughness, references, and novelty. Also extensibility. It should teach and be a basis on which others build.
In general, there are several types: engineering, science, design, art, etc. (After Bill Mitchell) An engineering thesis describes a solution to a problem and is evaluated by whether it works and improves the art. A science thesis proves a theorem. Design is evaluated by critique and iteration by experts. For example, redesigning the signage at an airport ought to show how it works better and is elegant as determined by peers in the field.
Note that when choosing a model it is tempting to pick a thesis that is not the best. If you want a model, chose the best. I have supervised over 50 theses, many stand out. Sometimes the thesis is the full picture of a student's graduate career, other times it is just one component. Discuss your model with me.
The most important thing about a thesis is that it be done. The statistical mode of thesis readership is 2—you and me (A paper gets a far larger audience.) We have broad latitude in what goes into it. (A PhD thesis, on the other hand, is perfect; time is the free variable.) You have to define a hypothesis that can be shown/proven in the allotted time. "In the fullness of time" is not an evaluation. E.g., I don't endorse a thesis that is based on the inevitability of VR.
Remember, a thesis is not a memoir. Your personal journey can be an organizing thread but not a theme. Posterity doesn't care about you—it cares about what you contributed.
Finally, all of these students left a positive imprint on the lab. Every student in the group or lab did better as a result of their contributions.