storage exception in ocr containers

Question

customer is experiencing a storage exception issue in their OCR containers. They are using the container version 3.2-model-2022-04-30 on Amazon Linux. The container is online. When calling the syncAnalyze endpoint, instead of getting analyze results, a storage exception is returned.

Can you enlighten me on how the container handles recognition results? Does the syncAnalyze API also store its results locally before returning the result?

If yes, can this behaviour be modified to not store any results locally?

Does the "StorageTimeToLiveInMinutes" environment variable affect how long the results are stored?

Thank you.