Handle historical (sampled) data properly

[dudebehinddude]

Currently, sampled data is just stored as an array in the program. This is a problem for multiple reasons: first of all, it is volitile, meaning that when the server is restarted all the data is lost. Second of all, it means that the data is slowly accumulating over time, increasing RAM usage. Additionally, with the current implementation, all the accumulated data is sent in the API, which is bad because that could potentially result in huge API calls. Here is a list of improvements that need to be made:

• Properly store and fetch historical data from a database
• Limit historical data calls to be from the past 1 hour/24 hours/1 week depending on the API call
• Add API calls to be able to fetch data from a certain date (in the past)