Inference on AWS Bedrock for OpenAI models has been very slow and glitchy the last few days with today being the worst. Very long latency of over 5 minutes to first token. What's going on?
us-east-1 seems to be the worst region right now.