Phase 0–1 operator burden review — East Tennessee two-site (hostile)¶

Stance¶

Default failure mode: the plan is under-budgeted on operator hours, not under-documented. Hidden burden is normal: travel, corrosion, battery anxiety, alert triage, restore drills, and “quick checks” from the phone compound while the family still has jobs, school, and weather.

This page does not add new gates (see Execution gates — Phase 0–1); it names load so EG passes are not confused with free time.

Burden lenses (what to evaluate)¶

Travel time¶

Recurring: ~70 min round-trip cab time per batch visit before productive work at SITE_FARM (Two-site operations model). Surge: bad weather, detours, trailer, stuck gate = P90 blows the spreadsheet.

Doctrine that helps: distance-tax rules, trip batching policy, P1 instrumentation ordering (water/gate class first).

Still thin: “two visits/week” as calendar truth without logged τ_ctx (gear, fuel, boots) and night emergencies outside batch windows.

Field maintenance¶

Recurring: fence, vegetation, culverts, lane surface on sloped ground (Demory site intelligence). Surge: storm washouts, escape repair.

Doctrine: site intelligence + capital sequencing peers earthwork with fence/water.

Thin: O&M $/hr for internal roads as a first-class line next to fence O&M (often lumped or ignored).

Off-grid power upkeep¶

Recurring: terminals, corrosion, BMS quirks, inverter warnings, genset policy if any, SOC discipline. Surge: prolonged cloud, heat, cold derates, equipment failure stacking with livestock stress.

Doctrine: Off-grid operational decision rules, Power domains and battery-backed tiers, degraded modes.

Thin: night response when telemetry says fine but physics disagrees—trust in dashboards vs mandatory physical rounds under stress is under-modeled emotionally.

Sensor upkeep and failure handling¶

Recurring: battery swaps, antenna aim, critter damage, firmware drift, false negatives (worst case). Surge: every sensor looks guilty during a welfare event.

Doctrine: observe-first Phase 1, stop rules NS family, single-pilot scope (Execution gates EG-SNS), sensor checklist matrix.

Thin: spares policy time (who drives which spare to which pole at 21:00) is rarely written until after first failure.

Gateway / controller upkeep¶

Recurring: power cycling stuck radios, SD corruption class failures, config drift, TLS expiry on edge. Surge: “gateway down” = blind until someone is on-site.

Doctrine: field-node classes, one RF class pilot, local-first / WAN-optional model.

Thin: two-site means two places things break—homelab patch night same week as 120 gateway swap is optimistic without naming who loses sleep.

Backup and restore obligations¶

Recurring: backup job monitoring, restore tests, credential rotation, ticket hygiene. Surge: actual restore during calving week.

Doctrine: Kubernetes platform backup/DR — Pi k3s Longhorn, backup cadence standards, EG-PLT gates.

Thin: “P0/P1 checklist green” is still hours and stress; families overweight “automated backup = safe” and underweight restore rehearsal labor.

Observability and alert triage¶

Recurring: Alertmanager routing changes, silences discipline, dashboard hygiene, log volume growth. Surge: alert storms during WAN/power transitions.

Doctrine: Observability, secrets, and trust bar, inhibition / welfare vs convenience split, CS-4.

Thin: On-call is often implicitly the same person who already drives 120 twice this week—no modeled collision between NOC hat and field hat.

Family labor reliability¶

Recurring: weekly rhythm fiction until V11 proves otherwise. Surge: illness, job travel, school events.

Doctrine: labor model falsifiers, coverage matrix, EG-LAB gates.

Thin: three-generation optimism—correlation of absence events is not independence; overlap weeks happen.

Weather / outage / degraded-mode workload¶

Recurring: seasonal adjustments. Surge: ice, mud, multi-day outage stacks WAN + power + water simultaneously (Risk register R16).

Doctrine: manual fallback SOPs, off-grid degraded modes, runbooks.

Thin: Drills are work; if drills are not scheduled, degraded posture is theoretical and burden shows up as panic.

Burden classes (summary)¶

Master burden table (Phase 0–1)¶

Severity: H = high (can dominate a week), M = medium, L = low if disciplined. Who: family role placeholders—replace with names locally.

Verdict (skeptical)¶

The narrowed Phase 0–1 scope is directionally right—observe-first, WAN-optional, single pilots, EG gates—but still optimistic if the family does not book triage hours, road minutes, power anxiety checks, and restore drills as first-class calendar items alongside field work.

Most under-modeled: collision between homelab ops nights and field surge weeks, and invisible phone triage that eats batching gains.

Related¶

• Labor model and weekly operating rhythm
• Risk register and mitigation strategy
• Observability, secrets, and trust bar — homelab / farm edge
• Automation stop rules — two-site smart farm
• Execution gates — Phase 0–1 (East Tennessee two-site)
• Instrumentation ROI model — two-site smart farm