Track and expose method to gauge local SDK instance's health

[ok300]

There are many things that can go wrong on a device, which may affect the SDK's ability to function.

A few ideas for what can be tracked:

• number of non-HTTP 200 API call responses + total number of HTTP API calls
• number of DNS failures
• number of API call timeouts

These could be tracked in a rolling window fashion, i.e. occurrences over the past 5 minutes, or since app was opened. They would ideally be tracked per endpoint.

We could expose the raw numbers, or an aggregate them in a simpler health metric (green / orange / red).

Once exposed, this could be used

• by asking it as an arg in the report_issue API
• generally by the app developer, such that if these numbers go too low, the app could show a warning health indicator and recommend solutions (check network, switch to WiFi, etc), temporarily guiding users away from starting new actions