Closed ThisIsMissEm closed 3 weeks ago
Whilst Retry-After is a Mastodon specific header
Um, no it is not: https://datatracker.ietf.org/doc/html/rfc9110#name-retry-after
I am happy to use whatever people are most comfortable with. I only had to implement this in a client once that I can remember and I am pretty sure this was with a Retry-After
header. I know other headers exist for a similar purpose, but went with what I knew here.
I also purposely omitted the Retry-After
header's other variant where it can give an HTTP-date
, because I think that number of seconds is easier to implement.
But if other headers are more popular and/or easier to implement in popular frameworks, then by all means we should use those.
Oh! I must've missed that, I'm guessing this could be why the aforementioned internet draft stalled.
In which case, I think we'd be well served to reference that RFC when talking about the header so everyone knows exactly what is to be expected
Whilst Retry-After is a Mastodon specific header, it would likely be best to suggest using the HTTP API Rate Limit header which is currently being defined (it's apparently "expired" at the moment, but I believe it's moving forwards): https://www.ietf.org/archive/id/draft-ietf-httpapi-ratelimit-headers-07.html