/fleet_health endpoint + transport-registry table

Sub-tickets

Questions

Activity

• system created · 2026-05-23
  wi cli; parent=#531
• system statusChanged · 2026-05-23
  claiming in parallel with #535 per PM. /fleet_health returns per-host snapshot: last_probe_ts, last_message_ts, latest_transport, agents_alive, agents_silent, status=green|yellow|red derived from silence_thresholds table. Same config table as #535; one schema, two consumers.
• system verified · 2026-05-23
  Shipped same v2.8.6. GET /fleet_health returns per-host {status: green|yellow|red, last_probe: {ts/transport/latency/ssh_r/zt}, agents: [{agent, last_seen, silence_s, threshold_s, status}]}. Status derived from worstFactor (silence/threshold) and probe age (>24h=red, >4h=yellow). Hosts: lezama/venus/whey detected; project-suffix agents with no probe land in 'unknown' bucket (informative, no alerts fire since rules don't match). Ready for coder-chatduo-cc to consume in chat-duo #537.
• system completed · 2026-05-23
  shipped + verified. GET /fleet_health endpoint serving per-host aggregated snapshot. Hub v2.8.6 same commit family. coder-chatduo-cc unblocked for #537 (Fleet Health pane).
• system completed · 2026-05-23
  shipped hub v2.8.6 commit 90c0d2c