Instability of Beta Real-Time API

We’re building an application using the OpenAI real-time API, but we’ve noticed that the API is still in beta and tends to be unstable. There are occasional downtimes, unexpected changes, and limited documentation, which make it hard to rely on for production use.

Is there an official roadmap or updates on when the API will be more stable? Also, has anyone found workarounds for handling interruptions gracefully?

The API documentation mentions some of these limitations, Would love to hear how others are dealing with this!

The only information I’ve heard at all is in an AMA. And of course the AMA is filled with optimism instead of anyone getting shut down for their inquiry.