Connection closed with error: [b'amqp:connection:forced', b"The connection was inactive for more than the allowed 240000 milliseconds and is closed by container '0d2b8a9ef31548df9b32154da33cc1cd_G21'.", None]

[breinlni]

• Package Name: azure.eventhub
• Package Version: 5.15.0
• Operating System: Linux
• Python Version: 3.11.4

Describe the bug
Hi there, we have a python service running that publishes some events to an event hub every few minutes. We're intercepting external logs and we're getting the following warning from time to time:

"Connection closed with error: [b'amqp:connection:forced', b"The connection was inactive for more than the allowed 240000 milliseconds and is closed by container '0d2b8a9ef31548df9b32154da33cc1cd_G21'.", None]"

Now, it does not seem like we have any problems with publishing events, so we assume that the connection is automatically re-created again in the background after closing it. Is that right?
We still would like to understand where this warning is coming from.
According to this documentation the idle timeout is "...by default None, meaning that the client will not shutdown due to inactivity unless initiated by the service."

To Reproduce
Steps to reproduce the behavior:

We are creating the producer client when we start our server (FastAPI))

producer = EventHubProducerClient(
    fully_qualified_namespace=settings.EVENT_HUB_FULLY_QUALIFIED_NAMESPACE,
    eventhub_name=settings.EVENT_HUB_NAME,
    credential=credential,
)

Expected behavior
If idle_timeout is not set, the connection should not time out

Screenshots

Additional context
We saw a couple of other issues related to that. But it's not clear how it answers our question:

• The connection was inactive for more than the allowed 240000 milliseconds and is closed by container #9917
• Unexpected error ErrorCodes.ConnectionCloseForced: The connection was inactive for more than the allowed 240000 milliseconds and is closed by container azure-iot-sdk-python#947
• ampq:connection:forced disconnect due to idle. Can't set idle time. Regression bug. #34212

[github-actions]

Hi @breinlni, I'm an AI Support assistant here to help with your issue. While the Azure SDK team reviews your request, I wanted to provide some possible tips and documentation that might help you in the meantime.

• The warning reflects that the connection was disconnected due to inactivity (“amqp:connection:forced”), which is expected behavior when no operations are performed within the service‐defined idle timeout.

  • The troubleshooting documentation explains that when idle, the service may drop the connection and the client will automatically re-establish a connection when an operation (like sending events) is attempted.
    • See the “ConnectionLostError Exception” section in our troubleshooting guide: TROUBLESHOOTING.md.

• In your scenario, since events are being sent sporadically, the connection closure due to inactivity does not impact event publishing because the client reconnects automatically when needed.

  • More details on client behavior and automatic reconnection can be found in the Azure Event Hubs Python SDK documentation.

The Azure SDK team will respond to your issue shortly. I hope these suggestions are helpful in the meantime. If this comment helped you, please give it a 👍. If the suggestion was not helpful or incorrect, please give it a 👎. Your feedback helps us improve!

[kashifkhan]

@breinlni this error is initiated by the service as described here https://learn.microsoft.com/en-us/azure/event-hubs/event-hubs-amqp-troubleshoot#connection-is-closed

Now, it does not seem like we have any problems with publishing events, so we assume that the connection is automatically re-created again in the background after closing it. Is that right?

yes the connection is recreated at that time as a brand-new Link has to be made. The SDK is reacting properly here as the service is not seeing any activity happen and closes the connection with that error.

[github-actions]

Hi @breinlni. Thank you for opening this issue and giving us the opportunity to assist. We believe that this has been addressed. If you feel that further discussion is needed, please add a comment with the text "/unresolve" to remove the "issue-addressed" label and continue the conversation.

[breinlni]

Hi @kashifkhan, so the event hub initiates the disconnect and the client automatically reopens the connection?
Is there any way for us to increase this timeout? Thanks for the help!