a365 setup all intermittently fails with ConnectionResetError (10054) during tenant setup

[EvanderDS]

Description

When running the Agent365 CLI command a365 setup all, the setup intermittently fails with a network connection reset error. The failure occurs at non‑deterministic points during execution and is not tied to a specific step.
This issue was observed while setting up a tenant for Agent 365 using the CLI. Similar failures have also been seen when running:

• a365  setup all
  

• a365 deploy
  

• a365 publish
  

In some cases, re‑running the command or restarting the machine temporarily resolves the issue, indicating a transient or environmental failure rather than a configuration error.

The issue occurs during the Agent 365 configuration and blueprint setup phase of the Agent 365 development lifecycle.

The following actions sometimes allow the command to succeed temporarily:

•    az config set core.disable_confirm_prompt=false
  

•    az config unset core.disable_ssl_verification
  

Expected behavior

1. a365 setup all should complete successfully and leave the Agent 365 environment in a ready state for the next lifecycle step (deploy / publish).
2. Transient network or TLS‑related failures should either:

•       be retried internally by the CLI, or
  

•       have clearer diagnostics
  

SDK Version

1.1.115-preview+fd1f775761

Language/Runtime

.NET 10.0.104 , Node.js v24.14.0

OS

Microsoft Windows 11 Enterprise

How to Reproduce

1. Open VS Code with the Agent 365 Toolkit installed
2. Open a terminal
3. Run az login
4. Run a365 setup all

https://learn.microsoft.com/en-us/microsoft-agent-365/developer/a365-dev-lifecycle

Output

The CLI fails with a connection reset error similar to:

        ConnectionResetError (10054): An existing connection was forcibly closed by the remote host

or:

      ERROR: Command failed with exit code 1:
     ('Connection aborted.', ConnectionResetError(10054, 'An existing connection was forcibly closed by the remote host', None, 10054, None))

Screenshots

Code of Conduct

• I agree to follow the Microsoft Open Source Code of Conduct.

[github-actions]

[AUTO-TRIAGE] Team Assistant Triage

This issue has been automatically analyzed and triaged.

AI Decision Reasoning: Type: Classified as 'bug' because the issue describes unexpected behavior where the CLI intermittently fails with a network connection reset error during tenant setup.
Priority: Assigned P2 because the issue disrupts the user workflow and requires manual retries, but it is intermittent and has temporary workarounds, reducing its criticality.
Copilot: This issue involves intermittent network connection errors and requires understanding of complex CLI behavior, retry logic, and potentially environmental factors. Copilot is not well-suited for debugging transient issues or implementing robust retry mechanisms without clear, scoped requirements.
Assignment: Grant Harris is a backend engineer with expertise in CLI, .NET, and Azure, which are directly relevant to the issue involving the Agent365 CLI and .NET runtime. Additionally, his contribution count is lower than the tech lead, making him a better choice workload-wise.
Labels: Applied labels bug, P2 based on issue type and priority

Labels and assignee have been applied based on this analysis.

[github-actions]

[AUTO-TRIAGE] Results

Classification: bug
Priority: P2
Confidence: 95.0%

Recommended Assignee: @gwharris7

Analysis:
type rationale: Classified as 'bug' because the issue describes unexpected behavior where the CLI intermittently fails with a network connection reset error during tenant setup.

priority rationale: Assigned P2 because the issue disrupts the user workflow and requires manual retries, but it is intermittent and has temporary workarounds, reducing its criticality.

copilot rationale: This issue involves intermittent network connection errors and requires understanding of complex CLI behavior, retry logic, and potentially environmental factors. Copilot is not well-suited for debugging transient issues or implementing robust retry mechanisms without clear, scoped requirements.

assignment rationale: Grant Harris is a backend engineer with expertise in CLI, .NET, and Azure, which are directly relevant to the issue involving the Agent365 CLI and .NET runtime. Additionally, his contribution count is lower than the tech lead, making him a better choice workload-wise.

labels rationale: Applied labels bug, P2 based on issue type and priority

[SUGGESTION] Suggested Approach

1. Investigate the retry logic in the CLI's network-related operations, particularly in the src/Services directory. Ensure that transient network errors like ConnectionResetError are handled gracefully by implementing exponential backoff retry logic using libraries like Polly for .NET.
2. Check the TLS and SSL configuration in the CLI codebase, especially if the CLI interacts with Azure or other external APIs. Review the configuration in the src/Services/AzureService.cs or similar files to ensure proper handling of SSL verification and connection settings. Consider adding an option to force TLS 1.2 or higher for secure connections.
3. Add detailed logging for network operations in the CLI to capture diagnostics when the error occurs. Update the logging mechanism in src/Services or the relevant service classes to include connection details, API endpoints, and retry attempts. This will help pinpoint the root cause of the intermittent failures.
4. Review the dependencies listed in autoTriage/requirements.txt for Python-based services that might interact with the CLI. Ensure that versions of requests, httpx, and aiohttp are up-to-date and compatible with the CLI's expected behavior, as outdated libraries might cause connection issues.
5. Update the CLI's documentation in the docs directory to include troubleshooting steps for transient network errors. Provide guidance on setting Azure CLI configurations (az config) and environment variables that can mitigate such issues, and include examples of retry mechanisms or diagnostic commands users can run.

---

Generated by TeamAssistant Auto-Triage powered by GitHub Models

[github-actions]

SLA Escalation Alert

This P2 issue has exceeded its SLA threshold.

Metric	Value
Priority	P2
SLA Threshold	72 hours
Time Since Update	84.1 hours
Current Assignees	@gwharris7

Action: Adding Tech Lead @sellakumaran as assignee for immediate attention.

cc: @ajmfehr

---

Generated by AutoTriage SLA Escalation