This release introduces smarter latency-based cluster selection, changes in configuration parameters, critical queue fixes, and updated documentation with new practical examples.
✨ Additions
- Latency mechanisms: automatically selects the fastest Edge Cluster Frontend if the MAX_LATENCY parameter is defined in requirements.
- Added example: usage of the MinIO service within an offloaded function.
- Every 600 seconds (if no requirements have changed since then), the device runtime will asked again for the Edge Cluster
- When stopping the device runtime device_runtime.stop() it will send a HTTP delete to
/v1/app_requirements/{app_req_id} - New user requirements added: ID (mandatory), PROVIDERS and IS_CONFIDENTIAL
- Internet connection is checked during the execution of a function to avoid the Device Runtime get stuck waitting for an answer
- Timeouts were added in all the HTTP requests
🔄 Changes
- The GEOLOCATION parameter in requirements is now mandatory and must include both latitude and longitude values.
🐛 Fixes
- CallQueue and SyncResultQueue are no longer Singletons, preventing concurrency issues.
- Fixed: Device Runtime gets stuck when the timeout triggerswhile executing a funcion
📖 Documentation
- Improved and updated documentation to reflect the latest changes and examples.