An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
Thank you for confirming that issue is mitigated from backend now.
Re-attached above as work around in case of temporary regional outages.Reduce full batch to mini batches
- Test in other regions (East US and Sweden Central are in demand regions)
- Leverage multiple batch deployment to reduce latency and throughput.
- Rest API commands were shared to cancel existing batch job if it is stalled for long. https://learn.microsoft.com/en-us/rest/api/azureopenai/batch/cancel?view=rest-azureopenai-2024-10-21&tabs=HTTP
We can setup monitor the events from log analytics workspace or set up a monitor in SDK.
if the processing time is more than usual processing time, it should execute above steps as precautionary measure.
Please take a minute to accept this answer if you found my recommendation helpful.
Thank you.