The API for document extraction in Azure Content Understanding returns an internal server error after 10 minutes of analysis when extracting tables that have more than 50 elements.

Question

The API for document extraction in Azure Content Understanding returns an internal server error after 10 minutes of analysis when extracting tables that have more than 50 elements.

Ilya 0

We created an analyser in Azure AI Foundry, with a schema set to extract rows of items from PDF invoices. When using the REST api to process some documents it works fine, unless the documents contain too many items (50+), where it crashes and we get this response data:

{

id: ‘***’,

status: 'Failed',

error: {

code: 'InternalServerError',

message: 'An unexpected error occurred.'

},

result: {

analyzerId: ‘****’,

apiVersion: '2024-12-01-preview',

createdAt: '2025-04-15T14:02:23Z',

warnings: [],

contents: []

}

kothapally Snigdha 2,185 Reputation points Microsoft External Staff

2025-04-16T04:51:17.9466667+00:00

Hi @Ilya

It appears that a significant number of items in the documents may be exceeding the memory or compute limits of the analyzer. To address this, we recommend adjusting the schema to reduce the complexity of the extraction rules. This can be achieved by focusing on extracting only the essential fields and simplifying the data extraction logic. Additionally, if supported consider batching or chunking larger documents into smaller, more manageable segments before sending them to the analyzer. Another potential solution is to configure the analyzer or API to process items in smaller batches (e.g., processing 25 items at a time instead of all 50+). Lastly, breaking the processing into parallel tasks by dividing the document into logical sections and processing them concurrently could reduce the load on the analyzer and improve overall performance.

I hope these helps you. Thank you!
kothapally Snigdha 2,185 Reputation points Microsoft External Staff

2025-04-17T04:56:04.3933333+00:00

Hi @Ilya

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.
kothapally Snigdha 2,185 Reputation points Microsoft External Staff

2025-04-18T01:31:11.8066667+00:00

Hi @Ilya

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

Your answer

kothapally Snigdha 2,185 Reputation points Microsoft External Staff

2025-04-16T04:51:17.9466667+00:00

Hi @Ilya

It appears that a significant number of items in the documents may be exceeding the memory or compute limits of the analyzer. To address this, we recommend adjusting the schema to reduce the complexity of the extraction rules. This can be achieved by focusing on extracting only the essential fields and simplifying the data extraction logic. Additionally, if supported consider batching or chunking larger documents into smaller, more manageable segments before sending them to the analyzer. Another potential solution is to configure the analyzer or API to process items in smaller batches (e.g., processing 25 items at a time instead of all 50+). Lastly, breaking the processing into parallel tasks by dividing the document into logical sections and processing them concurrently could reduce the load on the analyzer and improve overall performance.

I hope these helps you. Thank you!
kothapally Snigdha 2,185 Reputation points Microsoft External Staff

2025-04-17T04:56:04.3933333+00:00

Hi @Ilya

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.
kothapally Snigdha 2,185 Reputation points Microsoft External Staff

2025-04-18T01:31:11.8066667+00:00

Hi @Ilya

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

Share via

The API for document extraction in Azure Content Understanding returns an internal server error after 10 minutes of analysis when extracting tables that have more than 50 elements.

Your answer