Priority: HIGH for anyone using AI Gateway as an LLM proxy.
aiGatewayRequestsAdaptiveGroups has a sum.cost that attributes upstream provider cost (OpenAI/Anthropic/Gemini/Vertex) for traffic routed through AI Gateway — external LLM spend that neither Workers AI Neurons nor the CF bill can see.
- dimensions:
gateway, provider, model, cached, error, rateLimited, statusCode
- sum:
cost, erroredRequests, cachedRequests, uncachedTokensIn/Out, cachedTokensIn/Out
- Gotcha:
cost is $0 for the workers-ai provider (Neurons bills that separately) — combine with the Neurons monitor to avoid double-counting and blind spots.
Reference implementation (built + running in prod at Divinci): Divinci-AI/cloudflare-billing-kill-switch.
From FEEDBACK-from-divinci-deployment.md — real-world findings from the Divinci self-hosted deployment, 2026-06-17.
Priority: HIGH for anyone using AI Gateway as an LLM proxy.
aiGatewayRequestsAdaptiveGroupshas asum.costthat attributes upstream provider cost (OpenAI/Anthropic/Gemini/Vertex) for traffic routed through AI Gateway — external LLM spend that neither Workers AI Neurons nor the CF bill can see.gateway,provider,model,cached,error,rateLimited,statusCodecost,erroredRequests,cachedRequests,uncachedTokensIn/Out,cachedTokensIn/Outcostis $0 for theworkers-aiprovider (Neurons bills that separately) — combine with the Neurons monitor to avoid double-counting and blind spots.Reference implementation (built + running in prod at Divinci):
Divinci-AI/cloudflare-billing-kill-switch.From
FEEDBACK-from-divinci-deployment.md— real-world findings from the Divinci self-hosted deployment, 2026-06-17.