Observability 4: /boards, /map and home answer fast - boards 8-10 s and /map 34.7 s cold today #3693

Owner, 2026-10-20 22:00Z: "deep dive into forge observability … i want like grafana level stuff … great as btop, lazygit, lazydocker, k9s, vercel … tied to standards and close to the data so i can explore … observability should feel alive … a place with things that the crew thinks i should know … clean widgets to expand into boards and reports … crew driven. engaging and fun. gamified. but still high utility. high signal to noise. if its quiet, fine! dont just spam me to fill the box." Evidence: steward UI audit 2026-10-20 21:49Z (screenshots in ~/projects/forge-audit/ on the owner's Mac). Design brief: symptoms over causes; summary → board → explore → raw row; live by streaming, never polling; standards at the seams; dense but readable (tabular numerals, ≥ 12 px); quiet by default.

Done: /boards first byte < 300 ms (today 8–10 s every load, Server-Timing app≈7900–9700), /map cold < 1 s (today 34.7 s, layout;dur=34001 on a cold request, 2.6 MB), home < 300 ms (one run 9.2 s, kpis;dur=1032). Precompute or cache by data version, stream the shell first, and keep Server-Timing names. Test: budgets asserted on fixtures sized like production, relative to a same-run baseline (never wall-clock in loaded cells, see #3069).

Live triage (automatic, not written by a person): no pillar named, so it waits behind the Later gate; name a pillar to move it.

Parked by the crew: the check failed in the place it ran, not in the change, and failed the same way when it was run again (pull #3710 closed). Planner: split this smaller, or rewrite what done looks like; a person can also say what to try.

Steward (trevato): pull #3710 was parked by an infrastructure fault, not this work - the egress cap from #3719 stopped every check's package install (02:20-03:12Z). The forge is rolled back and the cap is fixed (c7ea1935). Reopened and re-checked; this strike does not count.