Following the design commit df96690.
- Two columns (owner, 2026-10-14): sender_check_days.messages (migration
0009, not deployed) holds the messages deferred mid-message, from
cleanup's warning; deferred stays the connections smtpd turned away.
AddSenderCheckDeferred takes which; the counter passes cleanup's as a
message. The card has "Connections turned away" and "Messages deferred";
it warns on either within 24 hours (one mark, as before).
- Supervisord silent at start: postfixStarted asks again every 2 s for 30 s
(postfixState: started, running, known). With no answer the downtime is
the time since the alive mark - an upper bound - booked with estimated = 1
(sender_check_days.estimated, MAX over a day); the card prefixes such a
period's time with "≈ " and its note says what ≈ means. Before, such a
start counted nothing and the number the owner decides by read low.
- Help topic sender-check and the log topic word for word as the mockup
(help.go).
Tests: store (connections and messages apart, a period with an estimated
day); logtail (cleanup's warnings are messages); cmd/panel (supervisord
silent: the time since the mark, estimated); view fixture and the Health
page at four columns with ≈; e2e reads the row's two counts (panel down:
connections; panel hung: messages). Schema at 0009 with the columns.
schema-migrations row 9, CHANGELOG [Unreleased], audit plan lines ticked
(two), roadmap 28/62.
Evidence: go vet ./... clean; go test of store, view, cmd/panel ok locally,
logtail but for the Windows-only test; the server run follows before the
push.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mixandClaude Opus 5.5
ac985d03a8