fix: cloud provider detection with running event loop

[Yaminyam]

Avoid calling asyncio.run() when an event loop is already running to prevent runtime errors. Detection is skipped at module load if a loop is active, setting current_provider to None.

[Copilot]

Pull request overview

This PR fixes a runtime error that occurs when the identity module is loaded while an asyncio event loop is already running. The fix prevents calling asyncio.run() from within a running event loop by checking for an active loop and skipping cloud provider detection in that case.

Key Changes

• Modified cloud provider detection logic to check for a running event loop before attempting to call asyncio.run()
• Sets current_provider to None when a running loop is detected, deferring cloud detection to avoid runtime errors

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The logging message "Detected environment: on-premise setup" is misleading when current_provider is None due to a running event loop. In this case, the cloud provider detection was skipped (not performed), not that an on-premise environment was detected. Consider adding a separate log message to distinguish between "detection skipped due to running event loop" and "detected on-premise setup".

[Copilot]

When a running event loop is detected and current_provider is set to None, the subsequent call to _define_functions() (line 341) will configure the module to behave as if it's running in an on-premise environment. This means that cloud-specific functionality (like instance ID detection) will fall back to hostname-based behavior even if the code is actually running in a cloud environment. Consider documenting this behavior or providing an alternative mechanism to detect the cloud provider after module initialization for cases where the module is loaded within a running event loop.

Suggested change

	# to avoid "cannot be called from a running event loop" error
	# to avoid "cannot be called from a running event loop" errors.
	# In this case, current_provider remains None and will be treated as
	# an on-premise environment by downstream logic. If accurate cloud
	# detection is required in such environments, callers should invoke
	# detect_cloud() from within the running event loop and update
	# current_provider accordingly.
	log.info(
	"Skipping automatic cloud provider detection at import time because "
	"an asyncio event loop is already running; defaulting to on-premise "
	"behavior until current_provider is explicitly set."
	)