Add retry_policy to activity Info

[ytaben]

What was changed

Add retry_policy to activity.Info class.
My understanding is that the information already comes in the proto message and the only downside is 'from_proto' call + memory.

Why?

We have had an increasingly common request to provide monitoring and alerting only on the final activity failure.
As it stands - the existing activity failure metric is reported on each attempt with no information to determine whether the failure is 'final'.

Note: While it is in theory possible to make a workflow interceptor - we would like the metric to be attributed to the actual activity service worker. It makes it easier to, for example, implement automated canary testing for the activity worker service that only uses final failure metrics.

We are also not able to take the approach of 'pass this info in activity args' since we want a framework centric approach.

Checklist

1. Closes

2. How was this tested:
   Unit tests passed locally
   TBD

3. Any docs updates needed?
   IDEs should reflect the change automatically on upgrade