Hi everyone,
Has anyone else recently seen a sharp increase in DataverseSearch (formerly RelevanceSearch) storage usage in their environments?
In our case:
Starting in July 2025, DataverseSearch storage in Production grew from ~3.5 GB to ~18 GB in just a few weeks.
Our Test environment (a recent clone of Production with identical configuration, quick-find views, and searchable columns) is only ~5 GB.
We are not using Copilot Studio, and we haven’t added any files or Dataverse knowledge sources.
Microsoft support says this is due to changes in how Copilot search data is indexed and stored, but I can’t reconcile this with the difference between our Prod and Test usage. Apparently this changed happened around July 7th.
Questions for the community:
Are you seeing similar sudden spikes in DataverseSearch usage?
Have you been able to identify specific tables, fields, or features that are driving the increase?
Have you found any effective ways to reduce the storage without losing important functionality?
Any insights, telemetry, or experiences would be greatly appreciated.
Thanks,
Alen
Hi Allen,
Thanks for sharing your experience—this is a very relevant concern.
Yes, Others Are Seeing It Too
You're not alone. Several organizations have reported a sharp increase in DataverseSearch storage starting around early July 2023, likely tied to backend changes in Copilot indexing, even if Copilot Studio isn’t actively used.
Production vs. Test Discrepancy
The difference between your Prod and Test environments could be due to:
Storage Optimization Tips
Microsoft Guidance
Let me know if this helps, and please mark the response as helpful if it answered your question 😊
Best regards!
Mansi Soni
243
Super User 2025 Season 2
Muhammad Shahzad Sh...
158
Most Valuable Professional
Abhilash Warrier
95
Super User 2025 Season 2