feat: add temporary field to default error

[zepatrik]

Related Issue or Design Document

This feature allows to easily tell the client which errors it should retry and which need intervention.

Checklist

• I have read the contributing guidelines
  and signed the CLA.
• I have referenced an issue containing the design document if my change
  introduces a new feature.
• I have read the security policy.
• I confirm that this pull request does not address a security
  vulnerability. If this pull request addresses a security vulnerability, I
  confirm that I got green light (please contact
  security@ory.sh) from the maintainers to push
  the changes.
• I have added tests that prove my fix is effective or that my feature
  works.
• I have added necessary documentation within the code base (if
  appropriate).

Further comments

[aeneasr]

I'd inverse this and say "no_retry" because this error might not even make it to the caller if something else fails (e.g. internal cluster network error).

[zepatrik]

No retry feels very weird, as you have to say NoRetry = false to specify it is retryable. But you want the default to be retry right?

[aeneasr]

Yes, default should be retry for most 5xx errors (e.g. 502 bad gateway) and and some 4xx errors (e.g. 429 too many requests).

Usually, one would use the Retry-After to indicate when the retry can happen again. However, most of the infrastructure currently doesn't support setting this value correctly.

To disable retry explicitly, one could set Retry-After to a date far in the future, for example.

In summary, I don't think setting this as a JSON key is an appropriate fix as it is very application specific. We should probably work with default HTTP headers that are understood by most clients already.

For the console, we can narrow down the codes that need retry, e.g. 429, 503, 502, 500

[zepatrik]

OK, closing as this fix is too specific.