#5740 "A metric that departs from its normal is recorded as an anomaly to
review, and is findable after it ends", milestone 513.
- Every minute each learned key's last 15 minutes is judged against its
hour's normal: above the usual hold x1.25 and by the metric's floor opens
an occurrence, which ends when the level comes back.
- metric_anomalies holds one row per item to review; a repeat while it
still waits joins it as an occurrence (metric_anomaly_spans).
- An ended row nobody reviews within anomalies.review_days (14, Settings ->
Thresholds) lapses; an ongoing one never does.
- Learning skips every hour an occurrence covers unless its anomaly was
marked normal: no normal is learned without review.
- Alerts -> Anomalies lists those to review first, then the reviewed and
lapsed under day dividers; marking is operator/admin and audited.
- The Overview widget and get_overview warn of open ones and count those
to review; the host page shows its own; the weekly report lists the
week's and the lapsed; MCP list_anomalies and review_anomaly.
- Reviewed and lapsed rows are pruned at data.retention_days, and kept
in backups like alert history.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#5507 "Alert emails through the outbox, gathered for 2 minutes"
(milestone 501). Replaces the in-memory core/alert_burst.py.
- outbox.hold_alert: an alert joins each recipient's held alert email, or
opens one due 2 minutes later (System 156 ruling). A Postgres advisory
lock keeps two alerts at once from opening two windows for one reader.
- email_outbox_events.alert_json (migration 0043) keeps the alert as it
fired, so the email written when the window closes says what it said.
- write_held writes it then, in the reader's zone: the usual alert email
for one alert, the list for several (outbox.alert_words).
- after_settle records on each AlertEvent, once every email carrying it
has settled: email_sent, or email_error naming each reader it didn't
reach and the server's reason.
- A held email survives a restart and goes out when its window closes.
- A refusal of a message's one recipient no longer repeats the address.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#5465 "One email per burst of alerts". When a switch reboots, every host's
agent-offline rule fires at once and each account got one email per rule.
Alert email now waits 30 s from the first alert: each recipient then gets one
message listing every alert in the window, oldest first, in their own zone
("[Steward] 3 alerts: 2 firing, 1 resolved"). One alert alone keeps its
usual email. Webhooks still post per alert, at once. Every alert's event
records the email's outcome. A burst held when Steward stops is not sent;
the alerts stay on the record and on the Alerts page.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Step #5429 "Reset by email: request, link and new-password page" of
milestone 497 "Passwords: change your own, and reset it by email".
- Sign-in page: "Forgot your password?" while a mail server is set.
- /forgot takes a username or email and answers the same either way; only
an active local account with a real address gets mail. Requests count
per address on their own throttle, so they never lock out sign-in.
- password_resets (migration 0039): a one-time key, stored as its sha256,
working for an hour; asking again replaces it.
- /reset/<key>: the new password, then signed straight in; every other
browser is signed out (sessions_since). Bad keys count as invite keys do.
- Email "[Steward] Reset your password", in the account's own zone.
- Audited: user.password_reset_requested, user.password_reset_by_email.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The operator's ask: an invite link sent by mail, the key in the URL, and
the new user choosing their own username, password and email.
- user_invites (migration 0037): the key's sha256 only, the address, the
role and days the admin chose, who sent it, used/revoked.
- core/invites: create (revokes an older open invite to the address),
renew (new key, new clock), find_pending, the four states, signup checks.
- /invite/<key>, outside sign-in and throttled like login: the invitee's
username, email (prefilled, theirs to change) and password; the account
takes the admin's role, the key is used up in the same transaction and
they are signed in. A bad key gets one plain message.
- Settings › Users: Invite by email (address, role, 1/7/30 days), sent from
the admin in their zone; without a mail server, or if sending fails, the
link is shown once to copy. Invites not yet used are listed with Resend
and Withdraw.
- Audit: user.invited, user.invite_resent, user.invite_revoked, user.joined.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#5335 "Emails Steward sends read as designed messages: an HTML alert email
and weekly report, with a plain-text twin".
- core/mail.py renders each email as multipart/alternative from
templates/email/: a table-and-inline-style HTML part and a plain-text
twin carrying the same facts. Colours are the design system's light
theme, resolved from theme.css; nothing is fetched when a mail opens.
- Alert email: state badge, a headline sentence ("Disk used on web01 is
94%."), rule, resource, reading, the condition in words, and the time in
the zone Notifications was saved in with UTC beside it; "View in
Steward" when Steward's address is set. Subject "[Steward] Firing: <rule>
— <resource> 94%". Alerts now carry their operator.
- Weekly report: monitors down, active alerts, the week's alert events,
uptime and busiest Traefik routers as small tables with links. Sent from
the configured From address; the "Generated … UTC" line and the stale
repo footer are gone.
- condition_text() states a condition from a rule's parts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
bvandeusenandClaude Opus 5.5