kpi-burn-alert.sh cannot see unregistered codex agents - read v_usage_resolved, not usage_events
kpi-burn-alert.sh reads usage_events directly and filters agent_name NOT IN ('', '(unresolved)'), so the burn alarm is STRUCTURALLY BLIND to every unregistered codex agent - the 137.7M venus codex tokens with no roster claim cannot trip it, and edge hosts (lezama, whey) store '(unresolved)' for ALL codex rows at rest, so their codex spend is invisible to the alarm entirely. A guard that cannot fire on the population it exists to watch is not a guard. Matters now because Elazar has shifted fleet work onto codex seats with an explicit order to monitor for wasted tokens. Fix: read agent_telemetry.v_usage_resolved (already in production, yours, already read by fleet-tokens.sh and token-usage-web.sh) instead of usage_events. Force it red after: confirm a known unregistered codex agent's spend now appears in the alarm's input. Measured by bin-venus-cc 2026-09-13; carved out of the cc stop because it serves Elazar's monitoring order.
Activity
-
Retired 2026-09-24 (Elazar order via nw-venus-ca, plan temp-delete-plan v2.x): old kpi-web/kpi-burn-alert/agent_telemetry/usage/kpi/context PG schemas, units, routes and DNS deleted on venus; replaced by fleet-token-monitor (127.0.0.1:9731, kpi.pensanta.com + token-monitor.pensanta.com). Obsolete, closed.