v2.0a17

This release introduces transient error retry with exponential backoff in the HTTP layer, refactors device attribute initialization, expands the known cluster URL registry, and updates getting-started documentation to clarify authentication prerequisites.

New Features

• Retry with Exponential Backoff
  • request_url now automatically retries on transient transport failures (DNS errors, connection failures, timeouts, proxy errors) with exponential backoff up to 3 attempts
  • Added RETRY_MAX_RETRIES, RETRY_INITIAL_BACKOFF, RETRY_BACKOFF_MULTIPLIER, and TRANSIENT_TRANSPORT_ERRORS constants to base.py
  • All other exceptions are raised immediately as ResponseError without retrying
• Unknown Base URL Warning
  • _resolve_base_url now validates the resolved URL against known cluster URLs
  • Issues a non-fatal UserWarning if the URL is not a recognized Central cluster, expected for Central-On Prem & other non-production environments

Improvements

• Added explicit self.materialized = False initialization at the top of Device.__init__ for clarity
• from_api branch in Device.__init__ now raises ValueError if device_attributes is None, preventing silent failures
• Replaced manual key-rename loop and setattr calls in Device.__init__ with new _apply_api_attributes() helper method
• Device inventory fetch in Device.get() now uses filter_str="serialNumber eq {serial}" instead of the search= parameter to match updated API behavior
• Expanded CLUSTER_BASE_URLS in constants.py with changes to API Gateway Base URLs
• Removed unused _return_client_credentials method from base.py
• Updates to Authentication & Quickstart guide