From a personal experience standpoint, I feel like a poll / survey that pops up that I am forced to take about my mood would probably negatively impact my mood. You are basically asking people to stop what they are doing, reflect on themselves, and then report. That is a good amount of brain power and energy you are asking for, even if the survey is short and quick.
The act of reflecting on one's own mood in and of itself is extremely difficult and sometimes taxing to many people. On top of that, the results of such surveys are controversial at best. Many scientists feel that self-reported surveys about mood and "feelings" result in a data set that is not consistently accurate and possibly misleading.
Also, polls/surveys inherently give forced answers. The idea that you must fit in to one of them 5-10 sets of moods about a specific scenario / topic is not very realistic. People are complex and moods are extremely complex. People can feel both excited and apprehensive about something but surveys might not even have "apprehensive" as an option so the survey taker checks off "nervous" which is read entirely differently by an analyst.
Alternatively, if you are truly interested in trying to understand the happiness of your employees, create an anonymous suggestion box. The good old tried and true method of typing up a suggestion and placing it in a box is always a solid solution, but since you are in the tech sector you could also create your own version of the "WHISPER" app that allows employees to post thoughts and ideas anonymously. You would be surprised what people write in those boxes at their place of employment when they know it is anonymous!
If you feel you must, once you have a solid user base actually using your suggestion box app, you can then start putting up "scenarios" and ask people to comment on those. For instance, you can push a "special notification" to the app users asking them to give their suggestions about a specific event that recently happened, or a made or scenario, or even a person. This allows the employees to communicate in their own words which is much easier to infer inflection in and understand the basis of the individual writing the comment and why they are writing it versus a cold, hard checkbox.
re: reliability and validity of self reported assesments of mood/emotion
re: the taxing nature of recalling emotion or non-sensory experiences