Metrics Overview

Per-service metrics reference for the NVCF self-hosted control plane. Each linked page lists metric names, types, sources, descriptions, and the labels and filters that make the metric useful in queries and dashboards.

Control plane services

• NVCF API: request rates, response status codes, and log event counts for the NVCF API service.
• Invocation Service: HTTP request counts, durations, and invocation error metrics for the invocation path.
• ESS: template rendering counters and HTTP client metrics for the Encrypted Secrets Service.
• gRPC Proxy: client connection counts, NATS pipe health, gRPC worker session-attach latency, and HTTP RED metrics for the gRPC proxy.
• State Metrics Service: per-function instance count, stage durations, request latency, and function metadata.
• SIS/Spot: HTTP client metrics for the Spot Instance Service.
• Function Autoscaler: OpenTelemetry metrics emitted by the function autoscaler service.

LLM services

• LLM API Gateway: request and routing metrics for the LLM API gateway.
• LLM Function Invocation Metrics Report: end-to-end LLM invocation path report.
• LLM Request Router: request router metrics for LLM traffic.

Per-function containers

• Init Container: restart counts and termination reasons for function init containers.
• Utils Container: restart counts, termination reasons, and worker service response metrics for function utils containers.

Datastores

• Cassandra: client request latency, timeouts, authentication failures, and endpoint connection metrics.
• Vault/OpenBao: pointer to upstream OpenBao telemetry documentation.

See also

• Observability for logging, tracing, and overall observability configuration.
• Example Dashboards for reference Grafana dashboards.