Error ModelArts.4206 Occurred in Real-Time Service Prediction
Symptom
After a real-time service is deployed and running, an inference request is sent to the service, but error ModelArts.4206 occurred.
Possible Causes
ModelArts.4206 indicates that the request traffic on an API exceeded the preset threshold. To ensure stable service running, ModelArts limits the inference request traffic on a single API.
Solution
Reduce the inference request traffic on an API. If ultra-high concurrency is required, submit a service ticket.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.