Add main I2C sensor Retry-loop

[tyeth]

This is a result of having retries in the drivers/wrappers, and a desire for them to be faster (more interrupt like). We are therefore removing unnecessary delays while waiting to retry one sensor multiple times, and instead retrying the entire suite of sensors up to three times before failing that periodic reading event.

I was trying to add another driver with a long retry time, and this work was the result instead. See #564

This is unfinished, as currently it will evaluate all sensors again, instead only failing ones need be retried.

There will be a follow up PR to remove any unnecessary retries during I2C sensor data reads once this is merged.

[tyeth]

@brentru ready for review

[brentru]

I think this is good! It increases how robust the update() loop is.

I have two notes:

Performance: I know you noted it in the PR description, but I want to get to it in this PR. Let's mark failed reads and only retry the sensors that fail.

Code complexity: I think the update() loop is growing in its complexity and provided an idea below about encapsulation