Envoy AI Gateway server request duration histogram (OTel gen_ai semconv)

WithDescription('Generative AI server request duration such as time-to-last byte or last output token.') - note the gateway is a PROXY, so this measures the full round trip including the upstream model's own latency, and it is measured at the gateway rather than inside any engine. 14 explicit bounds in a strict doubling series from 10 ms to ~82 s; the ladder stops at 81.92 s, so a percentile above that is not computable from this histogram. Envoy AI Gateway main (envoyproxy/ai-gateway), read 2026-10-05. Metrics go through an OpenTelemetry MeterProvider (internal/metrics/metrics.go NewMeterFromEnv), not a bare prometheus_client registry: a Prometheus reader is ALWAYS attached, and a console or OTLP exporter is added additionally when OTEL_METRICS_EXPORTER is 'console', or when it is neither 'none' nor 'prometheus' and an OTLP endpoint env var is set. OTEL_SDK_DISABLED=true turns off the extra exporters but not the Prometheus reader. That means the metric names are dotted OTel names (gen_ai.server.*), which the Prometheus exporter renders with dots replaced by underscores and a unit suffix appended - so the scraped series is NOT the literal constant in genai.go. Names follow the OpenTelemetry GenAI semantic conventions, not any engine's private scheme.