Will failed Index runs generate Cost in in the Foundry when Document Intelligence is used?

Question

Will failed Index runs generate Cost in in the Foundry when Document Intelligence is used?

Jonathan Lenz 0

I implemented the Document Intelligence Skill in my Azure Search indexing process. To surpass the free 20 document limit I connected the skill set to a foundry resource. The indexer failed after about 400 documents due to one being invalid. Now I'm wondering if the cost will only be generated if the indexer completes successfully or if it is accumulated for all document until it failed? I don't see cost on the Foundry resource yet.

Anshika Varshney 4,345 Reputation points Microsoft External Staff Moderator

2025-12-12T16:56:56.5733333+00:00

Hi Jonathan Lenz,

Thank you for reaching out on the Microsoft Q&A.

Costs for Azure AI Foundry (and the underlying Azure AI Search index services) are tied to the billable resources you provision and run, not whether an index run succeeds or fails.

Service charges are based on the capacity you allocate the number of search units (replicas/partitions) and any AI enrichment or embedding operations you trigger rather than the outcome of a specific indexing job. You’ll incur charges for the Azure Search service and any other Azure resources used during the run (e.g., OpenAI embeddings) as long as those resources are active and processing. Microsoft Learn

A failed index run itself doesn’t magically refund or avoid costs Azure still had to perform the work up to the point of failure so you’ll likely see associated usage charges. The only exception would be if a preview tier or free quota specifically covers the activity.

For cost control, monitor the Search service capacity and token/embedding usage in Azure Cost Management or use budgets/alerts so you’re not surprised by unplanned indexing jobs driving charges.

Yes, failed index runs incur cost just like successful ones, because the underlying resources are still consumed during the attempt. It’s the resource usage that costs money, not the result of the run.

I Hope this helps. Do let me know if you have any further queries.

Thankyou!

Your answer

Anshika Varshney 4,345 Reputation points Microsoft External Staff Moderator

2025-12-12T16:56:56.5733333+00:00

Hi Jonathan Lenz,

Thank you for reaching out on the Microsoft Q&A.

Costs for Azure AI Foundry (and the underlying Azure AI Search index services) are tied to the billable resources you provision and run, not whether an index run succeeds or fails.

Service charges are based on the capacity you allocate the number of search units (replicas/partitions) and any AI enrichment or embedding operations you trigger rather than the outcome of a specific indexing job. You’ll incur charges for the Azure Search service and any other Azure resources used during the run (e.g., OpenAI embeddings) as long as those resources are active and processing. Microsoft Learn

A failed index run itself doesn’t magically refund or avoid costs Azure still had to perform the work up to the point of failure so you’ll likely see associated usage charges. The only exception would be if a preview tier or free quota specifically covers the activity.

For cost control, monitor the Search service capacity and token/embedding usage in Azure Cost Management or use budgets/alerts so you’re not surprised by unplanned indexing jobs driving charges.

Yes, failed index runs incur cost just like successful ones, because the underlying resources are still consumed during the attempt. It’s the resource usage that costs money, not the result of the run.

I Hope this helps. Do let me know if you have any further queries.

Thankyou!

Share via

Will failed Index runs generate Cost in in the Foundry when Document Intelligence is used?

Your answer