Intermittent connection dropping

microsoft / vscode-tools-for-ai

Azure Machine Learning for Visual Studio Code, previously called Visual Studio Code Tools for AI, is an extension to easily build, train, and deploy machine learning models to the cloud or the edge with Azure Machine Learning service.

Other

321 stars 91 forks source link

Intermittent connection dropping #2372

Open snjsnj opened 1 month ago

snjsnj commented 1 month ago

Expected Behavior

Working on Azure Virtual Desktop: When initiating a VS Code remote connection from an AzureML workspace compute instance, the connection should remain stable and functional throughout the development session. Files should synchronize and save correctly, ensuring that no work is lost due to unexpected disconnections.

Actual Behavior

Unfortunately, the remote connection occasionally breaks without any visible error messages within VS Code. Developers may continue coding, unaware that their files are not being synchronized and saved properly.

Steps to Reproduce the Problem

Open VS Code on an AzureML workspace compute instance.
Establish a remote connection to a target environment (e.g., a remote server or container).
Begin coding or editing files.
Observe whether the connection remains stable or if it unexpectedly disconnects.

Specifications

VS Code Version: 1.89.1
Platform: Windows

tbombach commented 1 month ago

Hi @snjsnj, I apologize for this issue that you are encountering and I appreciate the details you included in this issue. I have a few questions to help me understand better:

You mentioned that you are using Azure Virtual Desktop. Are you using AVD to connect to an Azure VM, then establishing a VS Code connection to a compute instance from that VM?
How do you initiate your VS Code connection to the compute instance? (For example, do you click the "VS Code" link from the Azure Machine Learning Studio?)

Again, I apologize for the problems you are encountering.

snjsnj commented 1 month ago

Answers:

You mentioned that you are using Azure Virtual Desktop. Are you using AVD to connect to an Azure VM, then establishing a VS Code connection to a compute instance from that VM?: No, I connect to an AVD using "Remote Desktop", on which I then load the AzureML Workspace through my Edge browser. I then start my compute instance. From my browser, I click "Edit in VSCode Desktop", which starts VS Code on the AVD. When the connection between the AVD's VSCode and the Compute Instance breaks, no error message is thrown, leading to work not being synced and saved.
How do you initiate your VS Code connection to the compute instance? (For example, do you click the "VS Code" link from the Azure Machine Learning Studio?): I click the Edit in VSCode-> Edit in VSCode Desktop