Occasionally, my team sees this error when playing our game along with the manifestations of those failed requests. We make connections out to one of our servers in AWS (us-east-1 in Northern Virginia if it matters)
Initially, I thought it could be us, so I checked all of the load balancer/server logs and didn’t see anything anomalous - the error rates were extremely low and didn’t correlate with these reports. We also weren’t under any kind of load and could make requests to the exact same endpoints successfully with other clients.
Today myself and a teammate managed to catch this in prod. Around 9AM PST we saw this error in the logs, as well as network errors failing to load parts of avatars, sounds, animations, etc.
Looking at the developer console, we saw these network requests all had the response code -1, and all seemed to take between 1 and 2ms. 1-2ms is barely enough time to attempt a connection, assuming that network traffic is actually done. We waited and switched to another server around the same time and it got better (though we still saw that issue once or twice.)
The status code and quick failure rate scream load shedding, DC network congestion (especially because all of the assets failing to load), NAT/internet gateway reliability issues, or security checks (like anti-DDoS measures). Is there any explanation for these errors? Do we have any way of improving the success rates of these requests?
Thank you for your time!
A private message is associated with this bug report