[QUERY] The session channel '{0}' cannot be found in connection 'out-connection{1}'. (MessagingEntityNotFound).

[ZuluLame]

Library name and version

Azure.Messaging.ServiceBus 7.17.1

Query/Question

We are using Function Apps with ServiceBus triggers on subscriptions with sessions. From time to time exceptions get raised with outer message:
"The session channel '{0}' cannot be found in connection 'out-connection{1}'. (MessagingEntityNotFound). ..."

I found definition and a set of instantiations of MessagingEntityNotFound in azure-sdk-for-net/Microsoft.Azure.ServiceBus, but nothing with this message. Internet is also silent about it. Information on what session channels are and how they work with sessions/locks/connections was difficult to find. I'd appreciate any hints.

the whole parsedStack is attached.

parsedStack.json

Environment

azurefunctions: 4.1039.400.25216

[github-actions]

Hi @ZuluLame, 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.

• Enable detailed client logging to capture context around the exception. This can help determine if the session channel is being closed or if there is a network disruption. For details on enabling logging with the Azure SDK, see Logging with the Azure SDK for .NET.
• Verify that the configuration for session handling is set appropriately in your Service Bus trigger. Since the error occurs intermittently, reviewing your session lock duration and how sessions are accepted may provide more insight into whether a session is expiring or being released unexpectedly.
• Review any transient conditions (such as network blips or connection resets) that might lead to an unexpected disposal of the session channel. Transient errors like these should be handled by the retry logic in your function.
• Monitor diagnostic logs (for example, console output via AzureEventSourceListener) to see if additional warnings or errors are logged prior to the exception. This can help validate whether the session channel was closed intentionally by Service Bus or due to an environmental issue. For information on creating a console logger, refer to AzureEventSourceListener.CreateConsoleLogger.

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!

[jsquire]

Hi @ZuluLame. Thanks for reaching out and we regret that you're experiencing difficulties. The error that you're seeing is not generated by the Service Bus client library, but rather the underlying AMQP protocol. A channel is a low-level primitive that allows operations to flow between the client and service. For Service Bus, this is abstracted away by the library we use for the transport, so the client itself has no insight into channels nor influence over them.

Tracing the source of the transport library, the only occurrence of that message is here, which seems to imply that the service sent a "receive session" frame for a channel that no longer exists in the connection. I can't say authoritatively, but my best speculation is that there was a network delay that caused the frame to be received after the client's "accept session" operation timed out.

To know for sure, we'll need to engage the folks who own the Microsoft.Azure.Amqp library. Depending on where that goes, we may or may not need to refer you to the Service Bus service team to investigate/explain service behavior. As a first step, I'll need to collect a bit more information. Can you help me understand the end-to-end scenario in more depth? Specifically, can you share code snippets that demonstrate how you're using the Service Bus clients in your application and what exact package versions are in use. (Azure.Messaging.ServiceBus is not associated with Functions and does not contain the trigger or other integration points)

I'd also like to understand the frequency that you're seeing these messages and any errors that are being observed in the Azure SDK logs around this one surfacing. Those details can be found in the Azure SDK logs that flow into your AppInsights instance.

[github-actions]

Hi @ZuluLame. Thank you for opening this issue and giving us the opportunity to assist. To help our team better understand your issue and the details of your scenario please provide a response to the question asked above or the information requested above. This will help us more accurately address your issue.

[ZuluLame]

Hey @jsquire, thanks for the effort analyzing the issue. That's very reassuring that we can count on your help.

The end-to-end scenario is that we have a FunctionApp, where the Program.cs creates, configures with ConfigureFunctionsWebApplication, builds and runs a HostBuilder,

Then function is defined like:

 [Function("SubscriptionApp_ProcessDevicePropertyNotificationTopicMessages")]
 public async Task ProcessDevicePropertyNotificationTopicMessages([ServiceBusTrigger(Constants.DevicePropertyNotificationTopic, Constants.ServiceName, IsSessionsEnabled = true)] ServiceBusReceivedMessage message, CancellationToken cancellationToken)
 {
     await _executor.ExecuteFunctionInSubcontainerAsync<IMessageHandling>(x => x.ProcessMessage(message), cancellationToken).ConfigureAwait(false);
 } 

The Function attribute comes from microsoft.azure.functions.worker.sdk\1.17.4\

On the non production environment it was reported from, the day I took the logs, we had 1600 messages where 2 had this issue, so it's an order of magnitude of roughly a promile.