August 25, 2026

Generated lab notebook from daily_summary, plan_journal, plan_delivery_log, and setpoint audit data. It is intentionally chronological and may include in-progress cycles before validation.

Planner Execution Ledger

Planner checkpoints can acknowledge that the active plan is still suitable without writing a new public plan ID. Those no-change decisions are part of the audit trail.

Delivery events

Plan writes

No-change acknowledgements

Pending

Event	Delivered	Type	Status	Note
End-of-day review and reset	00:16	MIDNIGHT	timed_out	Delivery recorded without a public plan write.
End-of-day review and reset (catch-up)	00:48	MIDNIGHT	timed_out	Delivery recorded without a public plan write.
End-of-day review and reset (catch-up)	01:19	MIDNIGHT	timed_out	Delivery recorded without a public plan write.
End-of-day review and reset (catch-up)	01:51	MIDNIGHT	timed_out	Delivery recorded without a public plan write.
End-of-day review and reset	02:22	MIDNIGHT	timed_out	Delivery recorded without a public plan write.
End-of-day review and reset	02:53	MIDNIGHT	timed_out	Delivery recorded without a public plan write.
Morning planning cycle	06:23	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle (catch-up)	06:54	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle (catch-up)	07:26	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle (catch-up)	07:57	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle	08:29	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle	09:00	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle	10:05	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle	10:37	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle	11:08	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle	11:40	SUNRISE	wrong_action	Delivery recorded without a public plan write.
Morning planning cycle	11:55	SUNRISE	timed_out	Delivery recorded without a public plan write.
Morning planning cycle	12:27	SUNRISE	timed_out	Delivery recorded without a public plan write.
Solar peak planning checkpoint	13:04	SOLAR_MAX	acked	Acknowledged: SOLAR_MAX 13:04 MDT: live solar 1014 W/m2 overshoots bias-corrected forecast peak (~745-754) by ~35%, but house is already on the governing plan's compliance-first hot/dry posture and it is working: relay truth VENT_COOL_FOG_ASSIST with 100% wet+fog authority, house VPD actively recovering (13:02 avg ~1.9 kPa -> 13:05 1.48, east 1.08/west 1.17). All Tier 1 moisture levers already at hot/dry aggressive end (fog_esc 0... Resolved 13:07 MDT.
Forecast deviation	13:51	FORECAST_DEVIATION	action_completed	Delivery recorded without a public plan write. Resolved 13:53 MDT.
Peak Stress	15:04	TRANSITION	acked	Acknowledged: Peak-stress checkpoint 15:04 MDT: conditions tracking. House 84.7F/1.10 kPa VPD inside served band (high edges 85.2F/1.19 kPa), controller IDLE with all actuators off, dew margin 9.6F healthy. VPD peaked ~1.32 kPa at 13:00 and declined two consecutive hours; day's stress window complete and residual afternoon heat physics-limited (outdoor 79.9F, exhaust-only box). Wet thresholds band-coupled and auto-tracking (engage 1.24 = vpd_high+0.05, all 1.44 = vpd_high+0... Resolved 15:07 MDT.
Decline	18:44	TRANSITION	action_completed	Delivery recorded without a public plan write. Resolved 18:58 MDT.
Evening planning cycle	19:44	SUNSET	plan_written	Wrote public plan iris-20260825-1944. Resolved 19:58 MDT.

🌆 Evening Cycle (7:58 PM) — `iris-20260825-1944`

Status

validated

Outcome score

3/10

Changed parameters

band_track_fraction, cold_vent_guard_delta_f, cool_exit_hysteresis_f, cool_stage2_exit_hysteresis_f, cool_stage2_over_high_f, direct_wet_stress_min_dew_margin_f, direct_wet_stress_vpd_margin_kpa, dwell_gate_ms ...

Reflection

Validating previous cycle: iris-20260824-0017

Previous hypothesis: MIDNIGHT plan for Monday 2026-08-24. Sunday closed at legacy score/compliance 60.6, essentially equal to the 60.8 seven-day average. Temperature and VPD diagnostics were both weak at 49.6% and 48.8%; heat dominated at 10.86 h, followed by 9.26 h VPD-high and 1.43 h VPD-low, with 0 h cold. Dew safety held with an 8.1F minimum margin and 0 risk hours. Attributable/graded compliance and reliable utility/cost fields remain unavailable, so no cost or controller-attributable claim is made. The completed SUNSET plan iris-20260823-1947 is graded 1/10, exactly matching the deterministic anchor. Its 4.4 h window scored 31.9, with 40.0% temperature and 25.7% VPD compliance. Low-side outcomes were safe—0 h cold and VPD-low, only 0.03 h each heat and VPD-high—but the live-edge recovery and stable corridor hypothesis failed causally and at the endpoint. At replacement, the house was 72.9F and 1… Result: Revised 5 → 3 after the deterministic anchor returned 1. Anchor guardrail scorecard: 75 guardrail events, 61 held, 14 dispatched, guardrail_penalty 3 — 81% of the plan’s moisture writes were held by the sticky VPD-high guardrail and did not reach the ESP32. My initial 5 credited the safe outcome (all measurable goals met to the boundary: vpd_high 0.00h, vpd_low 0.00h, dp_risk 0h, 12.9F dew, heat 0h, 129 gal); the anchor prices the materialization failure I under-weighted. The causal gap the anchor does not see: the held writes clamped to conservative-but-reasonable values and the house stayed in-corridor with 0.01h stress, so the process miss did not become an in-window outcome miss. But the plan’s core mechanism — band-coupled wet authority to close the late-night VENTILATE VPD-high moisture-assist gap — mostly did not materialize (14/75 dispatched), and that gap is re-emerging at the boundary (alert 10427, 00:15, VPD-high no moisture assist 100%). Revising to 3: safe outcome above the anchor, but the core mechanism mostly failed to deliver, which is a real causal partial-failure a process score alone understates. The replacement MIDNIGHT plan aligns its first segment with the sticky guardrail values (engage ~vpd_high+0.05, all-zone ~+0.25) to reduce held repairs. Score: 3/10

Hypothesis

Testing: Whether keeping wet authority past 19:00 (wet_cutoff_hour 22) with a gradually widening band-coupled engagement threshold (mister_engage vpd_high+0.05 → +0.20) prevents the late-night VPD-high persistence seen 08-25 23:00-01:00 without creating pre-dawn VPD-low stress. Expected outcome: Overnight 19:45-06:00 MDT: vpd_high_stress under 1.5h, vpd_low_stress under 0.5h, dp_risk 0h with min dew margin >=8F, zero heat runtime, 08-25 total water under 200 gal.

Conditions

Outdoor peak

91.0°F

RH minimum

20.0%

Solar peak

827.0 W/m²

Cloud average

40.0%

tonight mild and increasingly humid (outdoor RH 44→82%, VPD raw 1.5→0.30, cloud 64-100%, no heating needed); tomorrow repeats hot-dry pattern (91F / 20% RH / ~3.9 kPa corrected VPD)

Expected stress windows

vpd_highmedium · 2026-08-25T20:00:00-06:00 to 2026-08-26T01:00:00-06:00

band-coupled wet authority kept open: engage vpd_high+0.05 tapering to +0.10, all-zone +0.25->+0.30, duty 10->8%, wet cutoff 22, stress override active

vpd_lowlow · 2026-08-26T02:00:00-06:00 to 2026-08-26T06:30:00-06:00

taper engage to vpd_high+0.15->+0.20, all-zone +0.35->+0.40, duty 5%, resource sensitivity 0.5; controller DEHUM owns residual excess humidity

Parameter rationale

mister_engage_kpa1.0 → 0.99; house VPD 0.87 kPa (east 1.09) above dispatcher high edge; last night 23:00-01:00 VPD 0.83-0.99 had no moisture assist (alerts 10424/10425); value is vpd_high+0.05 evening, tapers to +0.20 by 05:30

late-night (20:00-01:00) VPD-high stress reduced from ~2-3h to under 1h

mister_all_kpa1.2 → 1.19; outdoor VPD 0.30-0.42 kPa corrected by 03:00-07:00; humidity hold expected, pre-dawn band-low risk; value is vpd_high+0.25 evening, tapers to +0.40 by 05:30

pre-dawn VPD-low stress under 0.5h without a zero-wet cliff

fog_escalation_kpa0.25 → 0.35; fog blocked outside 07:00-17:00; tomorrow 08:00 is 64F/72% RH, unconfirmed dry ramp risks repeat morning over-mist

no fog burst at 07:00 window-open before live VPD confirms the dry ramp

mister_water_budget_gal300.0 → 150.0; authority mode RESOURCE_OPTIMIZATION_ALLOWED, band error 0.00, dew margin 8.4F, in band; resets to 300 at the midnight waypoint

overnight mist runtime under 0.5h; 08-25 total water under 200 gal

Setpoints

Tuesday August 25

Primary crop-band changes:

Time	Values	Note
19:45	hyst 0.1975	evening_settle (immediate): house at 80.2F / 0.87 kPa house

Tactical tunable changes:

Time	Values	Note
19:45	s2 1.76; cool_exit 1.78; all_fans 0; dw_stress 0; engage 0.9870000000000001; all 1.187; pulse 30; gap 48.75; wt 1.4333333333333333	evening_settle (immediate): house at 80.2F / 0.87 kPa house

Wednesday August 26

Primary crop-band changes:

Time	Values	Note
00:00	hyst 0.1975	midnight_posture: outdoor at 66F/66% RH rising, cloud 0-5%;
03:00	hyst 0.1975	humidity_hold: deepest humidity trough (outdoor 62F/78% RH,
05:30	hyst 0.1975	pre_dawn: outdoor 59F/78-82% RH, VPD 0.30-0.39; hold the VPD

Tactical tunable changes:

Time	Values	Note
00:00	s2 1.775; cool_exit 1.78; all_fans 0; dw_stress 0; engage 1.0370000000000001; all 1.237; pulse 30; gap 52.5; wt 1.3866666666666667	midnight_posture: outdoor at 66F/66% RH rising, cloud 0-5%;
03:00	s2 1.775; cool_exit 1.78; all_fans 0; dw_stress 0; engage 1.087; all 1.287; pulse 30; gap 52.5; wt 1.3666666666666667	humidity_hold: deepest humidity trough (outdoor 62F/78% RH,
05:30	s2 1.775; cool_exit 1.78; all_fans 0; dw_stress 0; engage 1.137; all 1.3370000000000002; pulse 30; gap 52.5; wt 1.3666666666666667	pre_dawn: outdoor 59F/78-82% RH, VPD 0.30-0.39; hold the VPD

Changed secondary parameters:

Time	Parameter	Change
19:45	band_track_fraction	initial 0
19:45	cold_vent_guard_delta_f	initial 8
19:45	cool_stage2_exit_hysteresis_f	initial 1
19:45	direct_wet_stress_min_dew_margin_f	initial 8
19:45	direct_wet_stress_vpd_margin_kpa	initial 0.05
19:45	dwell_gate_ms	initial 240000
19:45	enthalpy_close	initial 1
19:45	enthalpy_open	initial -2
19:45	fog_escalation_kpa	initial 0.35
19:45	heat_hysteresis	initial 1
19:45	min_fog_off_s	initial 84
19:45	min_fog_on_s	initial 54.75
19:45	mist_backoff_s	initial 600
19:45	mist_max_closed_vent_s	initial 600
19:45	mist_thermal_relief_s	initial 90
19:45	mister_all_delay_s	initial 84
19:45	mister_engage_delay_s	initial 42
19:45	mister_water_budget_gal	initial 150
19:45	night_vpd_bias_kpa	initial 0
19:45	outdoor_staleness_max_s	initial 600
19:45	sw_dwell_gate_enabled	initial 1
19:45	sw_fog_closes_vent	initial 1
19:45	sw_mister_closes_vent	initial 0
19:45	sw_summer_vent_enabled	initial 1
19:45	temp_hysteresis	initial 1.78
19:45	vent_exchange_fraction	initial 0.3
19:45	vent_prefer_dp_delta_f	initial 3
19:45	vent_prefer_temp_delta_f	initial 4
19:45	vpd_watch_dwell_s	initial 60
00:00	direct_wet_stress_vpd_margin_kpa	0.05 → 0.1
00:00	min_fog_off_s	84 → 90
00:00	min_fog_on_s	54.75 → 52.5
00:00	mister_water_budget_gal	150 → 300
03:00	direct_wet_stress_vpd_margin_kpa	0.1 → 0.15
05:30	direct_wet_stress_vpd_margin_kpa	0.15 → 0.2
05:30	fog_escalation_kpa	0.35 → 0.3

End-of-Day Summary

Climate

Temperature

67.0–85.8°F; avg 76.4°F

VPD

0.59–2.06 kPa; avg 0.95 kPa

Relative humidity

51.3–77.9%

Interior Light Evidence

Crop DLI: unavailable
Reason: source_reading_missing
Provenance: legacy_invalid_exterior_proxy_plus_fixture_estimate
Validity: dli-validity-v1; 2024-01-01T00:00:00+00:00 → open
Qualified-light minutes and photoperiod remain independent of unavailable DLI.

Stress Hours

Heat stress (>85°F): 6.3h
VPD stress (>2.0 kPa): 11.3h
Cold stress (<55°F): 7.3h

Economics

Electric

Not available

Gas

USD 0.00

Water

USD 0.620

Total