Production Orchestration — High-Availability Agent Execution

I will build production-grade orchestration infrastructure for your agent system — covering Kubernetes-based worker deployment with auto-scaling, a distributed task queue for orchestration job management, checkpoint persistence for workflow resumption after failure, rate limiting across all LLM API calls, and a comprehensive monitoring stack for operational visibility. Agent orchestration at production scale requires infrastructure engineering beyond the framework itself — Kubernetes deployment for agent workers, distributed checkpointing for long-running workflow resumption, and rate limit management across concurrent orchestration jobs are infrastructure requirements that the orchestration framework itself doesn't provide.

The infrastructure covers Kubernetes deployment manifests for orchestration workers, auto-scaling based on queue depth, distributed checkpoint storage, cross-worker rate limit management, health monitoring, and an operational runbook.