Working on a POC running on Azure and trying to get familiar with DocumentDB especially in regards to best practices/configuration. In our POC we have a very simple JSON doc with weather information and working on services to pull data back. Here's the issue but not sure if it's our mechanism for returning data or if I need to increase the performance level of the database (didn't want to start here as it seems like throwing hardware at a problem which may be resolved in other ways).
The performance issue that I came across was regarding the indexing of dates. Our web service method GetByDate needs to be able to form a range query to bring back all of the documents within a range of dates. I added a range index to the ReadOn column to DocumentDB, which worked but at a significant performance hit. Apparently some people are storing their dates as integers, others are storing both the string representation of the date and the integer, but only indexing the integer.
So hoping to hear recommendations for either modifications for how we are trying to query the database or if the only answer is to increase the performance level/scale out our DocumentDB even though it only has 5-10k records in it.