The answer is yes, ethnography can be defined as ultimately holistic; however, to do so, ethnographers (within the field of anthropology) have to make a concerted effort to incorporate holistic methods into their studies.
As of the moment, there are three paradigms to ethnographic inquiries: the holistic, the semiotic, and the behaviorist paradigms.
In the holistic approach, ethnographers don't begin with a hypothesis; basically, they don't let personal biases and preconceptions color their ethnographic studies of any existing cultural system. The holistic method allows those ethnographers who use this approach the flexibility to let the data speak for itself. It also allows the ethnographic researcher to accept the relevance of all contextual data and not just the typical data considered pertinent to the traditional, analytical process. In this light, even historical data is considered applicable to the process. So, the holistic ethnographic paradigm permits researchers the ability to conduct a participant-driven process instead of a researcher-driven (and therefore more biased) process in the study of cultural systems.
The other two paradigms, semiotic and behaviorist, hold significant advantages in the field of ethnography. The structural semiotic approach seems to place great emphasis on how the codes and rules within a cultural system define people within that system. On the other hand, the social semiotic approach investigates how people within cultural systems manipulate the prevailing codes and rules to their own advantage. As for the behaviorist approach, ethnographers who use this paradigm tend to allow for the inclusion of external observers in order to mitigate the possibility of falsifications or unintentional biases on the part of native subjects.
To recap, ethnography can be defined as holistic only if historical and all available contextual data are considered in the final analysis of each cultural system. As to whether the holistic paradigm is better or whether a combination of all three paradigms will make for better ethnography, that may be a discussion for another time. For more, please refer to the links below.