SSE Client Errors since 25.01.2022

we are using SSE for sending real-time events from our Python WebService (Python 3.8+, aioHTTP, gunicorn). The Service has been deployed last on 9th of August 2021 and has been running without any issues since, but yesterday we encountered issues with our SSE connections.

HTTP Requests work, EventSources get an initial connection but immediately receive an error indicating the connection has been terminated. What is weird is that the service is showing users as connected even though they have lost connection.

After encountering the errors on our main service we deployed a minimal setup that sends a simple timestamp event to a basic HTML Page to isolate the issue. We also tried deploying it on Heroku with the same result.

At this moment we are not sure if there is an external problem (like AWS/Cloudflare) or if anything has changed with how we should deploy our service.

Any insights or help would be greatly appreciated.