local-time: the panel in the container's zone, named on Settings

Agreed by the owner on 2026-10-13 (the local day, no "UTC" beside a
time); the zone is the container's TZ in .env and Settings only shows it,
his "Да" of 2026-10-17. Mockups c205106.

Every time the panel shows is formatted in time.Local and says no zone:
the logs' Time and the DMARC Received columns, a message's timeline,
Overview's "Updated", Health's certificate and renewal, the filter lists,
DMARC reports and windows, the auto limit's last recalculation. Stored
instants stay UTC. The days counted are the zone's: delivery_log_days and
sender_check_days key their rows by the local date, the sender check cuts
downtime at local midnights (byDay), the log filters' days run from local
midnight to local midnight. Notification mail and the test message write
the zone's abbreviation after a time - they are read away from the panel.
DMARC's one-day window label stays the reporter's (UTC) day.

Settings box 05, Time zone, read-only: the name (UTC without TZ), its
offset, and a TZ the image cannot load said to leave the times in UTC
(time/zoneinfo_unix.go, initLocal) rather than named. TZ is passed by the
compose file ("${TZ:-}"), in .env.example and the guide's table; tzdata is
named in the image (supervisor's python3.13 brings it today). Postfix
takes TZ through import_environment, so mail.log's stamps are the same
zone, which closes the trap the roadmap wrote down: the tailer has always
read them in time.Local.

Tests: a TestMain runs view, handlers, store and cmd/panel in UTC whatever
the machine's zone (as logtail's); TestSettingsShowsTheTimeZone,
TestTimesAreTheContainersZone, TestByDayInTheContainersZone, a Moscow day
in TestQueryInboundSpamFilters; envdoc's pattern takes two-letter names.
e2e time_zone_is_the_containers recreates the stand with
TZ=Europe/Moscow: date says MSK, Settings names the zone and UTC+03:00, a
test message's page gives its time and its mail.log lines in Moscow's.
go vet + go test (26 packages) and make e2e on the test server for this
tree.

Docs: guide (§ Settings, the TZ row, the days of Health and the logs),
architecture, schema-migrations, help (word for word with the mockup),
CHANGELOG [Unreleased] § Added, roadmap row done with the decision.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
mix
2026-10-17 01:34:28 +03:00
co-authored by Claude Opus 5.5
parent c205106418
commit bdeb01681d
58 changed files with 484 additions and 117 deletions
+4 -4
View File
@@ -475,10 +475,10 @@ the hole it closed was the same.
cleanup process serves message after message, so each such warning is
counted as a message. The two are counted apart (the owner, 2026-10-14):
connections turned away in `deferred`, messages deferred mid-message in
`messages` of `sender_check_days`, a row per UTC day kept a year, and
`messages` of `sender_check_days`, a row per day of the container's zone (`TZ`) kept a year, and
Server › Health shows today, 30 days and 12 months in a column each. The
counts also see a panel that is running and not answering, which the mark
does not. Downtime is cut at the UTC midnights it crossed. A log read again
does not. Downtime is cut at the midnights of that zone it crossed. A log read again
after an unclean stop does not count again: the count and the place in the
log it reached (the file's fingerprint and the offset past the line) are
saved in one transaction, and while the tailer catches up a warning at or
@@ -712,7 +712,7 @@ lines are gone for good — rotated past fourteen files while the panel was down
or deleted — or that the tailer was not reading.
Whether the sweep earns its place is measured, not argued (audit 2026-10-12,
H3). `delivery_log_days` (migration `0007`) keeps, per UTC day for a year:
H3). `delivery_log_days` (migration `0007`) keeps, per day of the container's zone for a year:
`reconcile_closed`, the rows the sweep closed; `reconcile_missed`, those of
them whose final line - sent, bounced, or the message's expiry - was in
`mail.log` or `mail.log.1` when the sweep closed them - the tailer did not
@@ -1203,7 +1203,7 @@ The tables that carry the panel's model:
`inbound_spam_list` (the filter lists), `dmarc_reports`,
`dmarc_report_records`, `notify_state` and `notify_addressees` (what the
notifier has said and to whom, § Notifications), `sender_check_days` (the
sender check's downtime and the mail it deferred, per UTC day) and
sender check's downtime and the mail it deferred, per day of the container's zone) and
`delivery_log_days` (the queue sweep's measurements, § Log tailer).
Two processes write: the panel, on one connection, and `dmarc-ingest`, which