Some servers are suffering under regular load #6
Description
Perhaps periodic polling of a bogged-down CTFd instance isn't a good idea.
The current ctfd backend samples the response time and disables more expensive features (falling back to just the scoreboard, if necessary). This behavior should be applied to all backends. Perhaps also auto-increment the polling interval if an update takes too long?