Xqt added a comment.

I don't think that it is fundamentally incorrect. The maxlag parameter is a good enough prediction what will be the next replag. I guess it is a good strategy to use it instead of running against refusal of requests over and over again or (another strategy we used) double throttleling after each blocked request.

Anyway I would propose another strategy: All requests should always pass but fetch the replag and wait this time for the _next_ request.


TASK DETAIL
https://phabricator.wikimedia.org/T144023

EMAIL PREFERENCES
https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Xqt
Cc: Xqt, jayvdb, valhallasw, Aklapper, pywikibot-bugs-list, Multichill, Sethakill, Mdupont, JJMC89, dg711, Anomie, Legoktm
_______________________________________________
pywikibot-bugs mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/pywikibot-bugs

Reply via email to