Azure OpenAI Data Files API - Timeout errors blocking all file operations; request bulk file deletion

Question

Azure OpenAI Data Files API - Timeout errors blocking all file operations; request bulk file deletion

Maksim Iudin 45

We have a production-blocking issue with Azure OpenAI Data Files that needs urgent attention.

Our batch processing pipeline depends on the Data Files API to upload and manage JSONL files. This has worked without issues for months. Starting recently, every single file operation fails with a 40-second timeout. We can't list files, we can't delete files, we can't do anything. The files API is dead for us. The Azure Portal shows the same timeout when we try to manage files through the UI.

We're stuck. We can't even clean up old files to see if that helps, because the list operation itself times out. This is blocking production workloads.

Error Details

Unable to fetch files
Timeout of 40000ms exceeded.
Trace ID: c6b9371f-a211-42a9-8a77-3bdc6f54adaa
Client request ID: 9419939b-76af-41c7-83c4-50229b3ba566

What we've tried

REST API calls: timeout
Azure Portal UI: timeout
Different times of day: same result

Request

Please delete all files in our Data Files storage. We cannot self-service this because even listing files fails. Once resolved, we'll implement file expiration to prevent accumulation.

Resource Information

Subscription ID: bd2ac0ac-ca22-4b8f-a3e8-991b631648c7
Resource name: studytube-ai-sweden-central
Region: Sweden Central
Approximate file count: 100,000+
Date issue started: 28.11.2025

Sridhar M 2,525 Reputation points Microsoft External Staff Moderator

2025-12-01T20:08:46.12+00:00

Hi Maksim Iudin

please provide the details requested in private message

Aryan Parashar 3,380 Microsoft External Staff Moderator

Hi Maksim Iudin,

I understand that encountering issues with the Azure OpenAI Data Files API can be disruptive.

To assist you, please find below the instructions and code required to remove all files from your Azure OpenAI resource.

Before running the script, ensure that the necessary Python package is installed:

pip install openai

Once installed, you may use the following Python code to list and delete all data files associated with your Azure OpenAI resource:

from openai import OpenAI

# -----------------------------
# SOURCE RESOURCE (read/delete)
# -----------------------------
client = OpenAI(
    api_key = "<API_KEY>",
    base_url = "https://<RESOURCE-NAME>.openai.azure.com/openai/v1/"
)

# ------------------------------------------------
# LIST FILES
# ------------------------------------------------
files = client.files.list()
print("Found", len(files.data), "files in the resource:\n")

for f in files.data:
    print(f"• filename: {f.filename}")
    print(f"  id:       {f.id}")
    print(f"  purpose:  {f.purpose}\n")

# ------------------------------------------------
# DELETE ALL FILES
# ------------------------------------------------
if not files.data:
    print("No files to delete.")

else:
    print("Deleting ALL files...\n")

    for f in files.data:
        try:
            print(f"Deleting: {f.filename} ({f.id})")
            deleted = client.files.delete(f.id)
            print("  Deleted:", deleted)
        except Exception as e:
            print("  Failed to delete:", f.filename, "-", e)

print("\nDone.")

You can locate your API key and resource information in the Azure portal as shown below.
User's image

I tested the process on my end, and it worked successfully.

Let me know if you are still facing issues.

Maksim Iudin 45 Reputation points

2025-12-05T12:31:07.1433333+00:00
Hello @Aryan Parashar !

Thanks for the snippet. I've created a similar one, slightly more sophisticated to handle API failures and timeouts. It took me about 20 hours to remove all the files and everything is working again. Since we had about 300,000 files stored in Azure OpenAI, the API simply couldn't respond within the configured timeout range (I guess there is N+1 issue or something like that). However, the probability of success was very much dependent on the time of the day. At some point I started getting successful responses from the files API.

I would like to share my observations:

We didn't send expires_after for the files. I thought in this case they will be stored for 30 days (default setting), but apparently they are stored forever. Is this a correct behaviour?

The documentation says the limit for the total number of files for batch inference API is 10,000. However, we were able to create 300,000+ files. I guess some checks are missing.

For now, the issue is resolved. Thanks for your attention.

Kind regards,

Maksim.

Your answer

Sridhar M 2,525 Reputation points Microsoft External Staff Moderator

2025-12-01T20:08:46.12+00:00

Hi Maksim Iudin

please provide the details requested in private message
Aryan Parashar 3,380 Reputation points Microsoft External Staff Moderator

2025-12-04T07:06:14.9966667+00:00

Hi Maksim Iudin,

I understand that encountering issues with the Azure OpenAI Data Files API can be disruptive.

To assist you, please find below the instructions and code required to remove all files from your Azure OpenAI resource.

Before running the script, ensure that the necessary Python package is installed:

pip install openai

Once installed, you may use the following Python code to list and delete all data files associated with your Azure OpenAI resource:

from openai import OpenAI # ----------------------------- # SOURCE RESOURCE (read/delete) # ----------------------------- client = OpenAI( api_key = "<API_KEY>", base_url = "https://<RESOURCE-NAME>.openai.azure.com/openai/v1/" ) # ------------------------------------------------ # LIST FILES # ------------------------------------------------ files = client.files.list() print("Found", len(files.data), "files in the resource:\n") for f in files.data: print(f"• filename: {f.filename}") print(f" id: {f.id}") print(f" purpose: {f.purpose}\n") # ------------------------------------------------ # DELETE ALL FILES # ------------------------------------------------ if not files.data: print("No files to delete.") else: print("Deleting ALL files...\n") for f in files.data: try: print(f"Deleting: {f.filename} ({f.id})") deleted = client.files.delete(f.id) print(" Deleted:", deleted) except Exception as e: print(" Failed to delete:", f.filename, "-", e) print("\nDone.")

You can locate your API key and resource information in the Azure portal as shown below.

I tested the process on my end, and it worked successfully.

Let me know if you are still facing issues.
Maksim Iudin 45 Reputation points

2025-12-05T12:31:07.1433333+00:00

Hello @Aryan Parashar !

Thanks for the snippet. I've created a similar one, slightly more sophisticated to handle API failures and timeouts. It took me about 20 hours to remove all the files and everything is working again. Since we had about 300,000 files stored in Azure OpenAI, the API simply couldn't respond within the configured timeout range (I guess there is N+1 issue or something like that). However, the probability of success was very much dependent on the time of the day. At some point I started getting successful responses from the files API.

I would like to share my observations:

We didn't send expires_after for the files. I thought in this case they will be stored for 30 days (default setting), but apparently they are stored forever. Is this a correct behaviour?

The documentation says the limit for the total number of files for batch inference API is 10,000. However, we were able to create 300,000+ files. I guess some checks are missing.

For now, the issue is resolved. Thanks for your attention.

Kind regards,

Maksim.

Share via

Azure OpenAI Data Files API - Timeout errors blocking all file operations; request bulk file deletion

Your answer