We identified and resolved an issue that caused our Deploy APIs to respond very slowly when generating embeddings via OpenAI.
We've fixed the issue, and the APIs are now up and running regularly. No downtime has been reported during this investigation.