In recent weeks, there's been some significant usage spikes that have caused some timeouts (502 errors) as the servers scale to more instances to accomodate the demand. We're currently reviewing the server infrastructure to make any necessary changes to ensure smoother scaling.