An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
Welcome to Microsoft Q&A and Thank you for reaching out.
Thread‑attached vector stores in Azure AI Foundry follow a default expiration policy of 7 days after last activity (meaning the last time the vector store participated in a run). Once this expiration point is reached, the vector store enters an expired state, and any runs on that thread will fail because the store is no longer considered active. [How to use...soft Learn | Learn.Microsoft.com]
When a vector store expires, it is not automatically deleted. The object simply transitions to an expired status (a documented lifecycle state), but all stored chunks, embeddings, and associated file data continue to occupy storage. Expiration only disables functional use; it does not free disk or memory by itself. [Vector Sto...soft Learn | Learn.Microsoft.com], [How to use...soft Learn | Learn.Microsoft.com]
Cleanup does not happen automatically at the 7‑day expiration mark. Vector stores remain in the system still visible in the portal and still occupying storage until you or your system explicitly deletes them using the vector‑store delete operation. There is no guaranteed time‑based auto‑purge documented for expired vector stores. Manual deletion is the only reliable cleanup mechanism.
If a vector store expires and thread runs begin failing, the only fix is to:
- Create a new vector store,
- Re‑upload the same files, and
- Reattach the new store to the thread. This reactivates the file search tool and restores run functionality. [How to use...soft Learn | Learn.Microsoft.com]
Note:Expiration happens automatically after 7 days of inactivity.
- Expired vector stores remain stored and consume space until manually removed.
- They are not cleaned up by the system automatically.
- Expired stores cause run failures but do not disappear on their own.
I Hope this helps. Do let me know if you have any further queries.
Thank you!