503 service unavailable timeout when sync'ing 700+ messages on heroku

[stephenmkbrady]

Timeout when sync'ing 700+ messages on heroku and not locally Log: ''' at=error code=H12 desc="Request timeout" method=GET path="/chat/" host=peaceful-sea-58238.herokuapp.com request_id=34dbf6e6-1572-4599-8a1a-601d00582329 fwd="" dyno=web.1 connect=1ms service=30000ms status=503 bytes=0 protocol=https ''' I think this might be something to do with gunicorn's worker timeout, but I upped it already to higher than 30seconds in the command line: web: gunicorn pymatrix_client.wsgi --timeout 120 --log-file -

Will have to investigate further. Local testing will not catch this, so need to see if I can get tests running on server with deployment.

[stephenmkbrady]

Remaining issue:

• Sometimes the matrix server is under load, will have to find a more graceful way to fail
• Lower the placeholder default value of messages for login sync
• Add a message about the matrix server load and high number of messages could be an issue. Resolved:
• Fixed a couple of issues with login: always joining when logging in to room, stopped calling api with an empty token, it was sometimes failing here when token = "", it needs to be None or a string

Not resolved:

• Sync'ing 700+ messages in a single REST is a bit much, will leave it possible, but might put an info box on the login with lower default.

This won't be fixed fully without some js ajax magic.