Azure AI Search Storage Limits and Index Updates

Question

When an index is updated in Azure AI Search, the index size temporarily increases beyond its original size before stabilizing.

Does the storage limit per service (e.g., 160 GB for S1) apply to this temporary increased size, or does it apply to the final size after the indexing is complete?

If it applies to the temporary size, will the indexer fail with an error if the temporary size exceeds the storage limit?

Answer 1

Hi junichi ito,

You're right, when you update an index (for example, by adding new fields or reindexing documents), Azure AI Search can temporarily increase the index size beyond its current footprint. This happens because Azure Search doesn't immediately remove old index data; instead, it creates a new version of the index internally while keeping the previous version live. Once indexing completes and the old data is cleaned up, the storage returns to a stable state.

So yes, the storage quota (e.g., 160 GB for S1) applies to the total physical storage used at any point in time, including:

• Primary index
• Temporary overhead during reindexing or updates
• Internals like storage for analyzers, fields, and history

So, if the temporary growth pushes usage beyond the quota, the operation may fail with an error like "Index quota exceeded".

1. Estimate your current and projected index size before running a large update. Azure doesn't auto-scale search storage, so proactive management is critical.
2. Options to avoid quota errors:
   • Scale up your service tier temporarily (e.g., from S1 to S2 or higher) to increase storage limit.
   • Split your index into smaller indexes if scaling isn't feasible.
   • Use indexers with soft-delete strategies to reduce re-indexing overhead where possible.
3. After the update completes and stabilizes, you can scale back down if you no longer need the extra capacity.

https://learn.microsoft.com/en-us/azure/search/search-sku-tier
https://learn.microsoft.com/en-us/azure/search/search-limits-quotas-capacity
https://learn.microsoft.com/en-us/azure/search/search-performance-optimization

Hope this helps, if you have any further concerns or queries, please feel free to reach out to us.