smolnar82 commented on PR #1409:
URL: https://github.com/apache/knox/pull/1409#issuecomment-5727089106

   > Thanks Sandor, looks good to me. Now we audit all the authorization 
decisions along with the exchange minting, and the refactor into helpers makes 
the handle method more readable.
   > 
   > I had thought we'd want to audit all failures. I think for SOC (number?) 
or FedRamp all request failures would need to be audited. We don't necessarily 
need to comply with any compliance specs, but it would limit use. Perhaps that 
opens us up to flooding the audit logs on DOS attacks and we want to avoid it, 
though I'd think in an enterprise deployment you'd have to handle that already 
via other/external means.
   
   You're right that all request failures should be audited; limiting 
compliance use isn't worth it. I've added auditing for every rejection, 
capturing who made the invalid request and (where known) for whom, with a 
reason code on each. The flooding concern is handled: the reason codes let 
downstream processing aggregate or rate-limit, on top of the external 
protections an enterprise deployment would already have. Tests added for each 
case.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to