v1.2.0

Streaming objects flow,
Batched checks in one swift go,
Watch your limits grow.

Added

• Larger result sets with StreamedListObjects
  While ListObjects has a limit of 1000 results, the StreamedListObjects endpoint has no specific
  result limits. When using this API, a network connection will remain open until the server has no
  more results to return, or after a timeout of 3 seconds, whichever occurs first.

• Batched authorization checks with BatchCheck
  This endpoint allows you to perform up to 50 authorization checks at once in a single request.

• Automatic retry with smart backoff
  Network hiccups happen. The SDK now automatically retries failed requests using exponential
  backoff with jitter, preventing thundering herd problems while ensuring your requests
  eventually succeed.

  • Configure retry attempts with the httpMaxRetries parameter (default: 3, max: 15)
  • Set to 0 if you prefer to handle retries yourself
  • The backoff algorithm spaces out retries intelligently to avoid overwhelming your FGA instance

• Observability support
  Understanding what's happening in production just got easier. The SDK now supports OpenTelemetry
  for metrics, tracing, and logging out of the box.

  • Track authorization latencies, error rates, and throughput
  • Trace requests across your entire stack
  • Debug issues faster with correlated logs
  • Configure with the telemetry Client constructor parameter
  • See the observability example for a complete setup guide

• Internationalization
  Error messages can now be displayed in your users' languages, making debugging easier for
  global teams.

  • Configure with the language parameter (defaults to English)
  • Currently supports Spanish and English

Changed

• Cleaner imports for helper functions
  We've streamlined how you import helper functions. Instead of multiple import statements, you can
  now grab everything you need in one line:

  use function OpenFGA\{allowed, dsl, model, store, tuple, write, ...};

  To avoid confusion, previous release notes have been updated to reflect this change.

• Centralized exception handling
  All exception messages now come from a single source, making them more consistent and easier to
  maintain. This sets the foundation for better error messages in future releases.