Issues

• Integrate, document, and verify rate limiting

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#22 · Filter by author luciusrockwing opened on Sep 25, 2026

• Expose authenticated aggregate stats

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#21 · Filter by author luciusrockwing opened on Sep 25, 2026

• Reconcile non-streaming usage and terminal fallbacks

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#20 · Filter by author luciusrockwing opened on Sep 25, 2026

• Reconcile SSE usage without breaking streaming

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#19 · Filter by author luciusrockwing opened on Sep 25, 2026

• Apply per-model budgets atomically

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#18 · Filter by author luciusrockwing opened on Sep 25, 2026

• Enforce global RPM/TPM budgets with Safety margin

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#17 · Filter by author luciusrockwing opened on Sep 25, 2026

• Make admission abort-aware and shutdown-safe

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#16 · Filter by author luciusrockwing opened on Sep 25, 2026

• Admit and queue work with bounded concurrency

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#15 · Filter by author luciusrockwing opened on Sep 25, 2026

• Add opt-in queue-and-wait RPM/TPM limiting with per-model budgets

  enhancementready-for-agent
  ak-a-ra/nvidia-api-proxy#13 · Filter by author luciusrockwing opened on Sep 25, 2026