Description of the new feature / enhancement
Improve error reporting from exceptions: currently the internal GH exceptions are lost so it is difficult to know if they are transient, rate limits, etc. Maybe just a --verbose would be good enough here to log exception stacks when they would otherwise be converted to a user message.
Improve retry logic (with backoff) so that transient errors are less impactful in automation scenarios. For example, when attempting to create the PR, it would be better to try a few times before giving the user an error to have them take over. This is true all of the time, but especially in automation where the only recourse is to run the pipeline again.
Proposed technical implementation details
No response
Description of the new feature / enhancement
Improve error reporting from exceptions: currently the internal GH exceptions are lost so it is difficult to know if they are transient, rate limits, etc. Maybe just a
--verbosewould be good enough here to log exception stacks when they would otherwise be converted to a user message.Improve retry logic (with backoff) so that transient errors are less impactful in automation scenarios. For example, when attempting to create the PR, it would be better to try a few times before giving the user an error to have them take over. This is true all of the time, but especially in automation where the only recourse is to run the pipeline again.
Proposed technical implementation details
No response