test urgent with agent

OPS-39 agent-ops Canceled urgent

backup-whey.service failing since 2026-05-09 02:03

OPS-31 agent-ops Canceled high

one-db: consolidate all whey telemetry/KPI sqlite into one local Postgres 17 db (telemetry), schema-per-domain, SELECT*-friendly views, reporting web UI. Goal: one place for all KPIs/perf/context data + Elazar can actually find/read them.

OPS-2 agent-ops Canceled high

└─ one-db: lazy-load telemetry-inventory.md — single doc agents read on demand listing, per data stream: which systemd unit/timer/cron collects it, which script the unit runs, which table it writes, and which script reports to Elazar + which aro. Keep it the SSOT map of the whole telemetry pipeline.

OPS-3 agent-ops Canceled high

└─ one-db: views layer — every table gets enough pg views (latest / daily / report rollups) so agents query a view with SELECT * (never name fields). Goal: zero token waste guessing columns. Document view list in telemetry-inventory.md.

OPS-4 agent-ops Canceled high

└─ one-db: cutover (NO dual-write — Elazar dropped it 2026-06-06). Collectors write straight to pg via metrics_write (spools to JSONL on pg-down, so no data loss); backfill already loaded full history. Flip 5 collectors + netreport off sqlite, retire sqlite files once pg-proven.

OPS-6 agent-ops Canceled high

└─ one-db: kpi.pensanta.com web UI on whey — small read-only server (chat-duo-web model, single-file .mjs) that reads the SELECT* views and renders KPI pages, so Elazar has one habit-friendly place to glance at all KPIs and drill into any metric when something looks off.

OPS-5 agent-ops Canceled

STRUCTURAL: /loop self-/compact over-cap workstream. Proven (Elazar 2026-05-21, ~28-30k ctx): a /loop session self-/compacts via ScheduleWakeup prompt /compact. Replaces cap->park->relaunch with cap->ss->self-/compact->continue. Gated on a ccs.sh launcher decision (ScheduleWakeup exists only in /loop-mode sessions) — needs Elazar GO. Separate from v2.7 over-cap path

OPS-51 agent-ops Canceled

Per-ccs maintainer & script/systemd ownership registry

OPS-26 agent-ops Canceled

└─ Authority matrix: per script/unit, who maintains, monitors, modifies, notifies. Decide: DB-backed (bwi) vs file-backed (~/.claude/maintainer-agents.md) source of truth.

OPS-28 agent-ops Canceled

└─ Process for new scripts: who creates, who maintains, alert routing. Capture in maintainer-agents.md once authority matrix decided.

OPS-30 agent-ops Canceled low

ZeroTier transport redundancy: zt_whey_visible=0 across all 5435 probes -> single public-dns path. Bring ZT up (zerotier-cli listpeers, whey visible) as a real fallback route

OPS-83 agent-ops Canceled ↳ OPS-80

Scope a HARDWARE remote-reset path for dead-kernel recovery (autossh/userland watchdog cannot revive a dead kernel): IPMI/iDRAC if present, else smart PDU - only no-physical-access option

OPS-84 agent-ops Canceled ↳ OPS-80