The Coach page gets a conversations panel on the right of the thread. From
1280 px it docks beside the thread and slides open and shut; the reading
column keeps its width either way and only its centring moves. It is open
on a first visit and remembers the last choice per device. Below 1280 px it
is a sheet from the right. The toggle sits at the trailing edge of the top
bar, through a small page-owned slot the top bar now renders.
The panel lists every conversation grouped by recency, with a row menu to
rename in place (Enter saves, Escape cancels without closing the panel) and
delete with the existing undo toast. The undo window now survives the sheet
closing. New chat is the round button at the panel's bottom right, and
picking or starting a conversation keeps ?c= in the URL.
The gear in the panel header opens quick settings: a popover from md up, a
bottom sheet on a phone. It switches the answering provider (chain reordered
through the existing route), the model and reasoning level where the person
owns them, and shows what the Coach can see. Keys, the Codex sign-in and the
fallback chain stay in Settings > AI. The drawer's gear opens the same
overlay. The composer's + menu is gone.
Settings > Coach links to /coach?settings=data, which opens the overlay on
what the Coach can see instead of a rail that no longer exists.
The in-memory day-aggregate fake re-filtered every row for every day,
converting each timestamp to the profile timezone again, so a 1000-day
series took about five seconds and the graded pulse-status test hit its
timeout. It now groups by day and segment in one pass (4.8 s to 85 ms).
The medications list and the compliance summary refetch on every mount,
so a take or skip logged on another device shows on return (#316). The
nav link's hover / touch / focus intent and the route-commit preload
prefetch both reads too, so every visit through a link sent each of
them twice: the prefetch, then the identical request from the mount a
few hundred milliseconds after the first answer landed.
The mount now skips its refetch only when the cached data answers a
prefetch that went to the network for this visit: within the last 10 s,
and since the route last left /medications. A direct load, a return
with an older entry, or a quick return whose prefetch found the entry
still fresh keep the refetch on mount.
A Coach snapshot, a derived-metric batch or the Insights page fans out
into a dozen readers that each probe the same account's rollup coverage
at once. In production one MCP get_metrics call held 59 of these probes
open on the database together, each taking about 320 ms under that
contention, and the same burst repeated every 15 minutes.
Concurrent probes for one account now join the query already running.
Nothing is kept after it settles, each caller gets its own map, and a
measurement write drops the in-flight entry so a probe that starts after
a write always reads afresh. The registry lives on globalThis so every
server bundle shares it.
A document whose preview render fails for good (a PDF whose first page
renders blank) was queued by the boot-time backfill, refused, and queued
again at the next boot, without end.
The thumbnail job now keeps the failure with the document
(document_thumbnail_failures, migration 0374): the reason, the attempts
since the last success and when the newest failed. The backfill, both the
boot discovery and the per-account pass, leaves a document out after one
deterministic refusal (raster-failed, empty-render, pixel-cap) or three of
any other kind, until the document changes. A direct enqueue still renders
it, and a preview that is made clears the record.
A weekly or nightly backup job whose process died (a restart, a
container out of memory) stayed active until its two- or four-hour
expiry, and its retry walked the accounts in the same order: the account
that killed it has the oldest copy, so it led again, and every account
after it went without a copy for as long as that record stayed too
large.
Both passes now send with a five-minute pg-boss heartbeat, so a dead run
fails within minutes and its retry picks it up. Each pass records when it
started and finished every account (backup_pass_attempts, migration
0373); an account whose last attempt never finished goes behind all the
others. The backups page shows when the run in progress started and names
an account a run died under, for the weekly and the off-host card, and a
heartbeat timeout reads as what it means.
Since v1.39.6 the preview scanner parsed the JSON on its way into the
pieces, kept every section but the three bulk tables, and validated the
whole document at the end: a second copy of the record beside the
writer's own, inside the weekly job. On an account with a document vault
that is most of the record.
The writer now shows each section and each bulk row to an observer just
before writing it. The preview collects key uses from that and takes its
counts from the writer's own FullBackupCounts; nothing is parsed or kept.
An upload hands over the summary and key uses it has already validated.
The round trip pins the writer's counts and key uses to a full read of
the copy on an account with a row in every model, and a memory test holds
what the pass keeps once the producer is done.
Measured on 70,000 measurements, 3,000 mood entries and 40 one-megabyte
documents under --max-old-space-size=524: held after the producer
56 MB before, 3 MB after; peak heapUsed 398 MB before, 334 MB after.
The drill read only the alphabetically last object of the newest date, so
one account stood in for all of them. It now checks every object of that
date one at a time, records a failure against the account it belongs to,
keeps going, and pages once per failing object with its key and how many
accounts were checked.
A text column written before key ids existed holds the bare base64 of
nonce, tag and ciphertext. The key check read such a value as an encoded
Bytes column, decoded it once too often, and accused the v1 key of not
opening a copy it opens. The instance's stored AI key has exactly that
shape and is the shortest v1 value in every disaster-recovery copy, so it
was picked as the probe and refused every restore of an account with v1
values, even with the instance settings left out of the verdict.
A legacy value now reads as a string unless its decoded head is base64
text. Samples are kept per section and only the sections the verdict
counts are probed; a key is unreadable when none of its probes opens.
Binary values open with the label their column is sealed under, taken
from the encrypted-column registry, and are not probed when no label is
known. A preview stored by an earlier release is still read, and its
sample's layout is derived again from the value.
Day keys and wall-clock fields are computed per row in the analytics,
digest and personal-record loops, and each call ran
Intl.DateTimeFormat#formatToParts plus a resolvedOptions() allocation
to re-validate the zone. In a CPU profile of an Insights page load the
zone helpers were 17-19% of the busy time on the request thread.
Probe the offset once per UTC hour at both ends of the hour and derive
the wall clock arithmetically; an hour that contains a transition, an
instant outside 1971-2199 or an invalid date keeps the formatter path.
Zone validity is memoised per name. userDayKey drops from 9.3 to 1.5 us
per call, wallClockInTz from 6.4 to 0.5 us; a parity test checks the
result against Intl around every 2026 transition in ten zones.
probeRollupCoverage runs ahead of the dashboard snapshot, the analytics
slices and the comprehensive insights. It read every live row of the
account for SELECT DISTINCT type and counted every DAY bucket per type,
which the planner served with a sequential scan of measurement_rollups.
Walk the distinct types with a recursive CTE along the live index and
test coverage with EXISTS. On a copy of a long-lived account the probe
drops from 24-50 ms to under 2 ms and returns the same map; an
integration test keeps the old query as the oracle.
A production build loads src/lib/db.ts once per server bundle, and the
client was only parked on globalThis outside production. The running
process held three clients: about 28 MB of query-compiler state each on
a heap capped near 520 MB, and three pg pools each sized to the full
per-process connection budget. Park the client on globalThis in every
environment so each copy of the module resolves the same one.
The tracker posts to `${data-host-url}${collect path}`, and the collect
path is compiled into the operator's Umami build: /api/send by default,
anything else when the instance sets COLLECT_API_ENDPOINT. The tag
pointed data-host-url at the app origin and the proxy listened only on
/api/send, so a renamed endpoint sent every page view and event to an
app path that does not exist and got a full 404 page back each time.
Point data-host-url at /api/monitoring/umami and answer every path under
it with one proxy route. The route reads the collect path from the
configured tracker (cached for an hour), forwards only that path to the
Umami host, and answers anything else, or an unreadable tracker, with an
empty response. The admin test event posts to the same learnt path. The
send-URL resolver lives in one module instead of two copies.
A sample the app wrote before the origin marker existed still carries the
HealthLog row id as HKExternalUUID. A hit on one of the account's rows of
the same kind now leaves it out exactly, for Withings and import mirrors as
well as manual ones; either id of a blood pressure pair covers both halves,
since the app stamps both samples of a reading with one id.
The manual-mirror lookup read one OR clause per row on every batch, even
for an account without a single MANUAL row. The types that hold MANUAL rows
are now read once per import; a batch asks nothing for any other type, and
one time window per type for those.
Cycle samples carry the marker too and are now left out like the rest, and
a sample the export repeats (a correlation's records also stand on their
own) is counted once.
The comment claimed the app's own sync filters by the marker. It does not:
the app's echo check is HKExternalUUID plus the sample's authoring source
being the app itself, and the marker exists for the deletion path, where
the source is not visible. Say so, and say why the export import reads the
marker and the row id instead.
The insights comparison snapshot took its current and baseline pulse figures
from a reading AVG, and the Coach's data inventory told the model a reading
mean of the whole pulse history. Both now read the mean of the user's local
days, each the mean of its hours' means, through the day-mean helper. Every
other type keeps its aggregate.
The feature extraction, the signals of the day, the status cards, the graded
series, the Coach snapshot, the mood correlations and the report's trend arrow
each still averaged pulse over its readings, so a dense workout hour decided
the day and a workout day the window. They now read a pulse day as the mean of
its hours' means and a window as the mean of its days, through the day-mean
helper. Counts, extremes and spreads stay over every reading; every other type
is unchanged.
The doctor report's pulse average and day points, the correlation channels,
the illness correlation, the personal baseline and the Coach's tiered context
each folded pulse readings into a day by plain mean. They now take the day as
the mean of its local hours' means and a window as the mean of its days, in
the zone each reader buckets by. The day aggregates carry the weighted sums for
an hourly-mean type, so every caller reads one day value. Every other type is
unchanged.
A blood pressure reading is a <Correlation> wrapping its systolic and diastolic
records, and the marker can sit on the correlation instead of on each record.
The import only read it from a record's own metadata, so a marked correlation
would have come back in full.
A correlation's records are now held until it closes, then all kept or all left
out together: the marker on the correlation, before or after its records, or on
a record, applies. Tests cover each, the marker not spilling onto the next
record, and a self-closing record.
The dashboard summaries, the insights aggregate and the series stats strip
still took a plain mean over pulse readings for their 7-day, 30-day, prior-month
and all-time figures, while the year-ago figure and the chart already weighed
each day once, so the year delta compared two different statistics. They now
go through the day-mean helper: each day is the mean of its hours' means and a
window the mean of its days, in the frame each reader buckets its days by.
Count, min, max, the median and the spread stay over every reading; every other
type is unchanged.
Rows written before this change keep the reading mean: the boot backfill and
the freshness check fill only missing days, so nothing rewrites them. Migration
0373 recomputes the mean of every stored PULSE row from the readings, DAY as
the mean of its hours' means and WEEK / MONTH / YEAR as the mean of their days,
and leaves count, min, max and the sums alone. The writer stores the coarse
rows the same way, so a refold and the migration agree.
A day of pulse is the mean of its hours' means and a window the mean of its
days, but only the rollup tier and the chart knew it. The helper gives every
other reader the same statistic in both shapes it needs: in TypeScript over
readings it holds, and in SQL as a per-reading day weight whose sums give the
mean of the day values over any window. Every other type keeps its plain mean.
The iOS app writes manual entries into Apple Health, and mirrors Withings and
import rows into it, each stamped with dev.healthlog.app.origin = healthlog.
The server already holds those readings under their own source, so the same
sample arriving through an Apple Health export is a second copy: the series
shows it twice and the rollups count it twice.
The export import now leaves out a record carrying that marker. The marker is a
child <MetadataEntry>, read after the record's open tag, so a record is
committed at its close tag. As a fallback for manual entries mirrored before the
app set the marker, a record is also left out when a MANUAL row of the same
type and value sits within 2 s (same rule and tolerance as the sync path).
Withings and import rows are not matched by value and time: they reach Apple
Health only through the mirror, which always sets the marker.
The import summary says how many records were left out for this reason.
Fixes#1109
An endpoint that does not know reasoning_effort answers with an unknown
parameter error, which the JSON-mode probe read as a refusal of
response_format and switched that base URL to plain prompts.
A client that rebuilds a form for a retry picks a fresh random boundary,
so the raw bytes differ and a repeated document upload ran a second time
instead of replaying. The boundary named in Content-Type is replaced by a
fixed token before hashing.
Reasoning models on Ollama and similar servers think by default and can
spend the whole max_tokens budget before the visible answer starts, so
the call comes back empty or times out. The Local and OpenAI-compatible
chain entries now carry an optional reasoning setting: Default sends
nothing, Off sends reasoning_effort "none", Low / Medium / High send
their own value. It is stored on the entry in aiProviderChain, so a
chain can mix a thinking model with one that does not know the field.
- Every exported resolver binds the entry's value onto the client, the
connection test included, keyed on the instance type so the operator
provider, OpenAI, Anthropic and Codex never carry it.
- PATCH /api/user/ai-provider accepts localReasoningEffort and
compatReasoningEffort and GET reports them. A provider absent from
the chain holds its value on a disabled entry; restating Default on a
chain that was never customised writes nothing.
- PUT /api/insights/provider-chain accepts reasoningEffort on local and
openai-compatible entries, refuses it on any other, and keeps the
stored value when the body leaves it out. Its schema is now shared
with the OpenAPI registry.
- A Default request body is byte-identical to the one sent before.
Refs #1126
A watch samples pulse about every 5 to 10 minutes at rest and about every
5 seconds in a workout. The daily value was the plain mean over a day's
readings, so a workout hour of several hundred readings outweighed the other
23 hours: a day with one hard hour and 23 resting hours (60 bpm) read about
128 bpm, not 64.
A day of pulse is now the mean of its per-hour means, in the rollup DAY
bucket, the live per-day read the chart serves, and the series route's day
grain. Windows over several days weigh each day once instead of by its
reading count, so the bias does not return one level up. Reading counts,
min/max, and the regression sums still run over every reading. The chart
tooltip and the data table name the statistic, in all seven locales.
The statistic is limited to pulse; HRV, SpO2 and glucose keep the plain mean
(see src/lib/measurements/day-statistic.ts). Stored DAY rows are rewritten by
the next rollup recompute (admin recompute route), not on read.
Fixes#1107
Refs #1124. The thumbnail job opened no wide event, so every annotation
it made went nowhere: one instance ran it hundreds of times, left sixteen
PDFs without a preview, and the log never said so.
Each run is now a `job.document_thumbnail` background event. A run that
should have produced a preview and did not is raised to warn and names
the reason; a type that never has a preview stays at info.
Refs #1124. The rasterizer opened pdfjs without its standard font, CMap
and wasm directories and told it to use system fonts. The runtime image
has none, so any text pdfjs could not draw from an embedded program (a
non-embedded standard font, or an embedded one it cannot parse) came out
as nothing, and an image-only provider was sent the page background. A
workstation hides this because its own fonts stand in.
pdfjs now gets the directories from its own package, resolved at runtime
from the working directory, and Liberation Sans from that package is
registered as the canvas' sans-serif, serif and monospace faces for the
text pdfjs draws with a substitute.
A document whose every rendered page is blank now reports render-failed,
so the read records raster-failed instead of indexing a transcript of an
empty page as a success.
The account's own mood categories, tags and hidden set were written only
inside the restore's mood-entry branch, so a vocabulary set up before the
first entry did not come back. They are written whether or not the file
carries an entry.
A practitioner's note, a visit's reason, outcome and body site and a
vaccination's note travelled as stored ciphertext in a portable file, the
same defect the cycle and mood sections had. They now go through the
shared sealed-text helpers: readable in a portable file, sealed under the
receiving host's key on restore, ciphertext kept only where this host
opens it and otherwise reported as visitCiphertext or
vaccinationCiphertext. The practitioner address and phone ciphertext of a
disaster-recovery file gets the same check, since the key preflight now
leaves these sections to it.
The replay cell was keyed on (user, key, method, path). A client that sends
two different writes under one key got the first write's cached response for
the second, and the second was never saved. Native clients up to 1.1.0 post
the two halves of a manual blood pressure exactly that way, so the diastolic
value was silently dropped.
Each cell now stores a SHA-256 fingerprint of the request body (JSON with
recursively sorted keys, otherwise the raw bytes; scoped by method and path;
at most the first MiB is read). Lookup replays, or refuses as in flight, only
when the fingerprint matches. A different body, or a row written before the
column existed, runs the handler without claiming or caching and leaves the
existing cell for its own retries; a lost claim race re-checks the winner's
fingerprint the same way. The body itself is never stored.
Migration 0371 adds the nullable idempotency_keys.request_fingerprint column.
The OpenAPI parameter description, the client notes and the Home Assistant
guide say what a replay now requires. docs/ops gains a read-only query that
lists systolic readings left without their diastolic half, held to the
schema by an integration test.
Refs #1085. Logging a workout right after finishing it meant changing the
start first, or the form refused it as ending in the future. The start field
now opens blank; left blank, the workout is saved as starting the duration
before now. A start the person enters is validated as before.
The labels of the mood tags and mood categories an account created
travelled as the stored ciphertext in a portable file, so a restore onto
a host with another key refused the file or wrote labels nobody could
open. They now follow the cycle note: readable in a portable file, sealed
under the receiving host's key on restore, ciphertext kept only where
this host opens it and otherwise reported as moodLabelCiphertext.
The sealing helpers move to one module both sections share.
The row update (name, schedule, oneShot) committed before the course
write. When a concurrent course made the course write refuse, the client
got the refusal while the rest was already written. The schedule replace,
the row update, the pause era and the course write now commit together,
so a refusal leaves the medication unchanged. The rollup recompute runs
after the commit.
A portable file carried the day-log note, the sensitive-category envelope
and custom symptom labels as the stored ciphertext, so a restore onto a
host with another key either refused the file or wrote values nobody
could open. The portable file now carries them readable and the restore
seals them under the receiving host's key; a disaster-recovery file keeps
the ciphertext.
A file written before this still restores: each ciphertext value is
opened with this host's keys first, and one that does not open is kept
out of the row and reported in the skip summary as cycleCiphertext. The
key preflight leaves the cycle sections to that per-value check.
Marc-André Bombeck