Modify the 422 errors cache to don't get deadlocked

[z16aclav]

Currently, there is a cache (in GitHubEnvironmentService.java) for 422 errors that don't retry to call Github for the last 1000 errors. When it happends, a standard error message is displayed.

It can create some issues:

1. deadlocks
   _As the decision to call github or not is solely based on the entry, if the 422 problem is fixed on the Github repo side, the update on Github will still be blocked
   _If there isn't a lot of new errors, github will never be called again
   This can create a deadlock where a repo will never be updated while it could be

2. overflow
   On the opposite side, if there is too many different errors sent, the Github API rate limit could be overflowed

3. lose of information
   _the 422 errors from Github come with a code indicating what specific issue happened https://docs.github.com/en/rest/using-the-rest-api/troubleshooting-the-rest-api?apiVersion=2022-11-28#validation-failed
   _this code is lost in the generic 422 - noretry - error message
   _With logs rotation, in case of a deadlock, it becomes impossible to know the reason why Github sent a 422 (and so it becomes way harder to fix the issue).

---

To handle all this issues, i would suggest to:
_add a time system to the cache so we both don't overflow nor deadlock based on the amount of different errors happening
_when Github sends a 422 error, store the code sent by Github (list -> map) and add it in the standardized error message

What do you think about it ?