v0.6.1

What's Changed

• 🚀 Autoscaling: Each agent autoscales independently on the concurrent in-flight load its broker handles.
• 🛡️ Broker rate limiting: Cap concurrent in-flight A2A requests per agent, shedding excess (HTTP 429 + Retry-After, gRPC RESOURCE_EXHAUSTED) to protect an external dependency like an LLM quota. Composes with autoscaling - scale-up is suppressed once total in-flight reaches the cap.
• 🧹 Graceful pod termination: Draining before termination, smooth and precise autoscaling down and pod rotation.
• 📈 Broker observability.

Full Changelog: v0.6.0...v0.6.1

Installation

kubectl create namespace kynomesh-system

# Install kynomesh controller
kubectl apply -n kynomesh-system -f https://github.com/kynoproj/kynomesh/releases/download/v0.6.1/install.yaml