We are using a Twisted HTTP server https://twistedmatrix.com/documents/15.0.0/web/howto/using-twistedweb.html and we are experiencing a strange behavior.
This HTTP server serves a static file and in some cases it may returns a 403 to one person (tested in the office) while the rest of the team can download the file without any problem.
If the person receiving a 403 runs a curl the curl works fine, it looks like is something browser specific but my question is can a HTTP server error be cached for a specific request?
After a while the static file can be download fine by the user experiencing a 403.
Also how a 403 (forbidden error) can just affect to just a single request and not to everybody? note there is not any kind of authentication.
Edit: As far as I have tested, looks like the problem just appears in Google Chrome, with Firefox or curl it gets rendered fine although Chrome keeps giving me a 403. That browser is caching somehow the petition and if I try with incognito mode it works fine.