WHY: any single typed fetch failure was a failed pass, so the whole
board flipped unavailable on one dropped poll and stayed dark until the
API answered again — even though HA's coordinator machinery keeps the
last good data on a failed pass and the panel could safely keep
rendering it for a while.
The coordinator now caches every successful snapshot with its
time.monotonic stamp; while the elapsed time since the last success is
within the configured staleness threshold (CONF_SNAPSHOT_STALENESS,
seconds, default 3x the update interval = 900 s) a typed failure serves
that cache instead: the panel keeps rendering last-good data, entities
stay available, and the outage is logged as a clear warning. Past the
threshold — or before the first success — the pass fails as before
(typed error logged and chained into UpdateFailed); a later successful
fetch refreshes the cache and the entities recover with no config-entry
reload. The threshold rides the options flow (validated as a positive
int at least the update interval, with a dedicated error string) and
the strings files; the clock is read through a _monotonic seam so the
staleness tests drive it deterministically.
Notify target (empty default, household fills their own service),
three strict HH:MM automation times (08:00/15:00/20:00 — the report
deliberately before the midnight sweep), and four independent enable
toggles, all leniently stored and round-tripped.
One refresh pass builds the whole entity snapshot: active children,
today's instances with derived open/done/missed states, per-child
presence via the PresenceEngine (records converted to model objects),
rollup counts, and the custody cycle day. Interval configurable via
the new update_interval option, default five minutes.
Review round 6 (Codex gpt-5.6-terra, MEDIUM): HA replaces entry.options
wholesale, so a submission omitting the picker dropped the allowlist
copy and left only the stale config-flow copy in entry data — a later
database loss would resurrect removed admins. Omitted submissions now
carry the current DB allowlist forward on the options copy. The
legacy-path test updated to assert the carried copy. Suite: 910 passed.
Review round 3 (Codex gpt-5.6-terra, LOW): a duplicate raw submission
reached set_admin_ids and crashed the flow with an unhandled
ValueError; the flow now rejects duplicates as invalid_admin with a
test proving the database is untouched. Suite: 905 passed.
- admin_allowlist.py: list_admin_ids / set_admin_ids (atomic sync of
the allowlist inside one transaction under the connection lock;
refuses empty targets so the last admin can never be removed and
the allowlist can never be emptied — fail closed) and
seed_setup_admin (seeds ONLY an empty allowlist: stored entry-data
copy first — disaster recovery for a lost database — else the HA
user who completed the config flow from the entry context;
non-empty lists are never resurrected).
- __init__.py: both setup paths (first setup and the reopened-record
path) seed after migrations; a failed seed closes the database and
re-raises through the existing error handling.
- options_flow.py: the admin picker is a multi-select of EXISTING HA
users (hass.auth.async_get_users), defaulted to the current
database allowlist, membership-checked at the schema level and
backed up on the raw-input path; empty selections are refused
(no_admins), unknown ids rejected (invalid_admin); the selection is
applied to the DB when the runtime record is reachable and always
stored on the entry as the recovery copy. Submissions without the
key (legacy raw payloads) keep the three-key behaviour untouched.
- const.CONF_ADMIN_USER_IDS; strings + translations for the field and
the two errors; conftest entries gain a context dict (HA's entry
context carries the flow user).
- Tests: 20 allowlist business tests (round-trips, refusal matrix,
seed precedence and narrowed-list persistence), 5 picker flow tests,
3 setup-seeding lifecycle tests. Suite: 901 passed.
overseer
dev_agent
orchestrator-agent
Joshua King