Schedule poll/notify and the digest as systemd timers

news-cycle.timer runs poll then notify every 15 minutes; news-digest.timer runs
the briefing at 08:45 daily. Both enabled and verified: a manual cycle fetched
1047 items, embedded 95 new ones and clustered them in 74s.

Units are installed as COPIES rather than symlinks. The previous service's units
were symlinked into docs/projects/news/deploy/, and when that repo was renamed
to news_triage they became dangling and the timer failed every ten minutes,
unnoticed, until this rebuild found it. Copying costs a re-copy after editing;
symlinking cost days of silent failure.

Deliberate details: the poll ExecStart is `-` prefixed so a failed fetch does not
skip notify; RandomizedDelaySec spreads polling off the quarter hour rather than
hitting 56 sources at once; the digest is Persistent so a machine that was off
at 08:45 still sends rather than skipping the day.

deploy/README.md

@@ -127,12 +127,47 @@ sudo systemctl disable --now news-update.timer
 | `POST /api/feedback/.../bogus` | `{"error":"bad signature"}` | reaches the app, unauthenticated, and rejects a forgery |
 | `POST /api/feedback/.../<valid HMAC>` | `{"ok":true}` and a row in `feedback` | **the loop that never once worked in v0.2** |
 
-## 3. Scheduling
+## 3. Scheduling — APPLIED 2026-09-08
 
+Two systemd timers, units in this directory:
+
+| unit | cadence | does |
+|---|---|---|
+| `news-cycle.timer` | every 15 min | `poll` then `notify` |
+| `news-digest.timer` | daily 08:45 | `digest` |
+
+```sh
+sudo cp deploy/news-*.service deploy/news-*.timer /etc/systemd/system/
+sudo systemctl daemon-reload
+sudo systemctl enable --now news-cycle.timer news-digest.timer
 ```
-*/15 * * * *  podman exec news-backend python -m news.cli poll
+
-*/20 * * * *  podman exec news-backend python -m news.cli notify
+**Copy the units, do not symlink them.** The previous service's units were
-45   8 * * *  podman exec news-backend python -m news.cli digest
+symlinks into `docs/projects/news/deploy/`, and when that repo was renamed to
+`news_triage` they became dangling and the timer failed every ten minutes,
+unnoticed, for days. The cost of copying is remembering to re-copy after editing
+a unit here; the cost of symlinking is a silent failure the next time a
+directory moves.
+
+Design notes:
+
+- `news-cycle.service` has two `ExecStart` lines and the first is prefixed `-`,
+  so a failed fetch does not skip `notify` -- there may still be ripe clusters
+  from an earlier cycle worth sending.
+- `RandomizedDelaySec=90` on the cycle: feeds get polled by everyone on the
+  quarter hour, and this both spreads that load and avoids hitting all 56
+  sources at the same instant.
+- `Persistent=true` on the digest, so a machine that was off at 08:45 still
+  sends the briefing on the next boot rather than silently skipping a day.
+- `TimeoutStartSec=20min` on the cycle: a cold start embeds the whole backlog at
+  ~11 items/sec. Steady state is far quicker -- a real incremental cycle
+  measured 74s for 95 new items.
+
+Reading the logs is noisier than it should be, because podman writes its own
+container-exec events into the same unit journal:
+
+```sh
+journalctl -u news-cycle -o cat | grep -vE '^[0-9]{4}-[0-9]{2}-[0-9]{2} .* container '
 ```
 
 ## 4. Verifying it actually works

deploy/news-cycle.service

@@ -0,0 +1,15 @@
+[Unit]
+Description=news: fetch, embed, cluster, then send any alerts
+Documentation=file:///home/connor/docs/projects/news/deploy/README.md
+After=network-online.target
+Wants=network-online.target
+
+[Service]
+Type=oneshot
+# `-` on the poll so a failed fetch does not skip notify: there may still be
+# ripe clusters from an earlier cycle worth sending.
+ExecStart=-/usr/bin/podman exec connor_news-backend_1 python -m news.cli poll
+ExecStart=/usr/bin/podman exec connor_news-backend_1 python -m news.cli notify
+# A cold cycle embeds the whole backlog at ~11 items/sec, so it can legitimately
+# run for minutes. Steady state is far quicker; this is a ceiling, not a target.
+TimeoutStartSec=20min

deploy/news-cycle.timer

@@ -0,0 +1,13 @@
+[Unit]
+Description=news: run a poll/notify cycle every 15 minutes
+
+[Timer]
+OnBootSec=3min
+OnUnitActiveSec=15min
+# Feeds are polled by many people on the quarter hour; spread the load and stay
+# a good citizen. Also avoids every source being hit at the same instant.
+RandomizedDelaySec=90
+AccuracySec=30s
+
+[Install]
+WantedBy=timers.target

deploy/news-digest.service

@@ -0,0 +1,11 @@
+[Unit]
+Description=news: write the daily briefing and push a pointer to it
+Documentation=file:///home/connor/docs/projects/news/deploy/README.md
+After=network-online.target
+Wants=network-online.target
+
+[Service]
+Type=oneshot
+ExecStart=/usr/bin/podman exec connor_news-backend_1 python -m news.cli digest
+# One LLM call, which can take ~20s; allow for a slow upstream.
+TimeoutStartSec=5min

deploy/news-digest.timer

@@ -0,0 +1,12 @@
+[Unit]
+Description=news: daily briefing at 08:45 local
+
+[Timer]
+OnCalendar=*-*-* 08:45:00
+# If the machine was asleep or down at 08:45, still send it on the next boot
+# rather than silently skipping the day.
+Persistent=true
+AccuracySec=1min
+
+[Install]
+WantedBy=timers.target