ADF - Az function activity 499 error

Question

ADF - Az function activity 499 error

Santhi Dhanuskodi 365

WE get an error while executing azure function activity inside adf pipeline. Azure function is calling an API and extracts and stores the data returned by api to a storage account.

Call to provided Azure function 'api_blob' failed with status-'499' while invoking 'POST' on 'https://<fucntionapp>.azurewebsites.net' and message - 'Invoking Azure function failed with HttpStatusCode - 499.'.

Pls note that the data is generated and stored in storge account, even though adf throws error. have set azure funciton timeout as 30 minutes, and also adf az function activity timeout is also set as more than 1 hour...And we use premium only for function app...and the data generation happens from 6 minutes to 8 minutes.

SAI JAGADEESH KUDIPUDI 1,825 Reputation points Microsoft External Staff Moderator

2026-03-09T08:34:40.13+00:00

Hi Santhi Dhanuskodi,
Could you please share requested details in private message
Santhi Dhanuskodi 365 Reputation points

2026-03-09T08:50:50.53+00:00

what details needed
Santhi Dhanuskodi 365 Reputation points

2026-03-10T07:09:10.8733333+00:00

Can someone respond to this question pls
SAI JAGADEESH KUDIPUDI 1,825 Reputation points Microsoft External Staff Moderator

2026-03-11T15:06:24.39+00:00
Hi Santhi Dhanuskodi,
The HTTP 499 error occurs because Azure Data Factory (ADF) closes the HTTP connection before the Azure Function returns its response, even though the function successfully completes API calls, data processing, and storage writes. Since the Azure Function is invoked synchronously and runs for 6–8 minutes, ADF may terminate the connection due to internal gateway handling, which causes Azure Functions to log HTTP 499 and ADF to mark the activity as failed.
Even though:

Azure Function timeout is configured to 30 minutes

ADF activity timeout is configured to more than 1 hour

Function App is running on Premium plan

Increasing timeout values does not reliably fix this, and Microsoft recommends using an asynchronous pattern where the function returns immediately and processes the workload using Durable Functions or background execution.

To resolve this issue permanently, the Azure Function should be redesigned to follow an asynchronous execution pattern, such as Durable Functions or queue-based processing, instead of keeping the HTTP request open until processing completes.
Recommended Design is Queue-Based Processing

This pattern decouples ADF from the long-running process.

ADF → HTTP Function → Queue → Queue Trigger Function → API call → Storage

Uses:

Azure Functions

Azure Storage Queue

Azure Data Factory
Step 1 — Create Storage Queue

Go to Azure Portal

Open your Storage Account

Select Queues

Click + Queue

Name it

api-processing-queue

Step 2 — Modify HTTP Azure Function (Trigger Function)
This function only pushes a message to the queue and exits quickly.

[FunctionName("StartApiJob")]

Result:

Function returns HTTP 202 immediately

No long wait for ADF

No 499 error Step 3 — Create Queue Trigger Function
Create another function:

Trigger type → Queue Trigger

Example:

[FunctionName("ProcessApiJob")]

This function performs:

API call

Data processing

Save to blob storage Step 4 — Update ADF Pipeline
In Azure Function activity inside Azure Data Factory:

Call:

https://<functionapp>.azurewebsites.net/api/StartApiJob

ADF receives:

HTTP 202 Accepted

Pipeline succeeds immediately.

Hope this helps. If you have any follow-up questions, please let me know. I would be happy to help.

Please do not forget to "up-vote" wherever the information provided helps you, as this can be beneficial to other community members.
SAI JAGADEESH KUDIPUDI 1,825 Reputation points Microsoft External Staff Moderator

2026-03-13T19:43:09.2766667+00:00

Hi Santhi Dhanuskodi,
I hope you had a chance to review the information shared earlier, and I hope this information has been helpful! If you still have questions, please let us know what is needed in the comments so the question can be answered.

1 answer

Your answer

SAI JAGADEESH KUDIPUDI 1,825 Reputation points Microsoft External Staff Moderator

2026-03-09T08:34:40.13+00:00

Hi Santhi Dhanuskodi,
Could you please share requested details in private message
Santhi Dhanuskodi 365 Reputation points

2026-03-09T08:50:50.53+00:00

what details needed
Santhi Dhanuskodi 365 Reputation points

2026-03-10T07:09:10.8733333+00:00

Can someone respond to this question pls
SAI JAGADEESH KUDIPUDI 1,825 Reputation points Microsoft External Staff Moderator

2026-03-13T19:43:09.2766667+00:00

Hi Santhi Dhanuskodi,
I hope you had a chance to review the information shared earlier, and I hope this information has been helpful! If you still have questions, please let us know what is needed in the comments so the question can be answered.

Answer 1

Hi @Santhi Dhanuskodi
Since the Azure Function is completing the processing and writing data successfully, the HTTP 499 error is related to how the connection between ADF and the Function is handled, not a failure in your business logic.

In Azure Data Factory, the Azure Function activity makes a synchronous HTTP call and expects a response within the gateway handling limits. If the function takes several minutes to respond (even if timeouts are configured higher), the connection can still be closed by the underlying service infrastructure, which results in a 499 error.

This explains why:

Data is generated and stored correctly
But ADF marks the activity as failed

What you can try first

Ensure the function returns the HTTP response as soon as processing completes (avoid any delayed response after blob write).
Check if the function logs show long response serialization or network delays after execution.
Test invoking the function directly (Postman / curl) to confirm response timing consistency.

Recommended long term approach

If this process may continue to take several minutes, consider using an asynchronous pattern such as:

Durable Functions

Queue-trigger based processing

This avoids keeping the HTTP request open and makes pipeline execution more reliable.

Share via

ADF - Az function activity 499 error

1 answer

Your answer