Api: During maintenance, auth endpoints return 400 for server error

[bhollis]

During the Shadowkeep maintenance, I noticed that the auth endpoints were returning 400 with a payload of "SystemDisabled". I'd expect a 500-series error for this - in DIM, we assumed 400s meant that the auth token was not valid and logged people out.

[floatingatoll]

For bonus points, use 503 + Retry-After to indicate either # of seconds until next retry or outright “don’t try again until XYZ”:

503 Service Temporarily Unavailable
Retry-After: Tue, 01 Oct 2019 17:00:00 GMT

Most clients will ignore the retry-after, but any that want to stand up a countdown clock can then do so :)

On Sep 30, 2019, at 15:26, Ben Hollis [e-mail] wrote:

https://user-images.githubusercontent.com/313208/65921261-9bccca80-e396-11e9-9910-d866df081472.png
https://user-images.githubusercontent.com/313208/65921262-9bccca80-e396-11e9-953b-8498db2cefb3.png
During the Shadowkeep maintenance, I noticed that the auth endpoints were returning 400 with a payload of "SystemDisabled". I'd expect a 500-series error for this - in DIM, we assumed 400s meant that the auth token was not valid and logged people out.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub https://github.com/Bungie-net/api/issues/1048?email_source=notifications&email_token=AAAWUDDCYR2EG5PNCJLGN23QMJ4LBA5CNFSM4I4BQAG2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HOV5NXA, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAWUDEFLLBVGS5FFWHTAFDQMJ4LBANCNFSM4I4BQAGQ.