We have a process that hits the events API once per minute to poll for events since a known stream_position (i.e. since the last stream_position that we checked at). Generally, this is a very reliable process, but suddenly around 5:00 AM MDT on 6/24, we started receiving periodic 404s on GET.
There isn't a solidly predictable pattern, but we'll get between 10 and 30 minutes worth of successful GETs (again, we do this once per minute), and then out of the blue we'll get a 404...followed by another 10 to 10 minutes worth of successful GETs until the next 404.
Is anyone else seeing similar behavior?