Open duncte123 opened 2 years ago
Some of the API responses can be a little slow however, so what kind of time frame are we thinking? 10 seconds? 15 seconds?
I would say 30 seconds, but it also needs to take into account the 502 response code and not cache that
30, can do. I was thinking maybe a fused timer that pits the API response and just a 30s sleep call against each other to handle that. As for error responses, they should probably be given as the reason for the error instead of the timeout. This shouldn't be too bad to make. And, conveniently, I just taught my local API simulator how to throw errors.
The 502 does not come from the API, it comes from caddy (the HTTP proxy).
Does the system bypass caddy when querying the API from inside of Docker?
Currently, v2 takes ages to load if the backend can't be reached. There should be a quick fail showing an error page saying that Oengus will be back soon.