Backups

Three backup paths, each with its own GitHub deploy key (deploy keys are
repo-scoped, not account-wide). Coverage completed 2026-11-01 — all 14 modules'
code and all module databases are now backed up.

1. In-run DB archive

• Source: /opt/sportstlm/data/sportstlm.db (opened read-only).
• Output per run: inrun-*.sqlite.gz (3 inrun tables, ~1.65 MB) and
  fulldb-*.sqlite.gz (whole DB, ~6.9 MB, 37 tables). Row counts + integrity verified
  before push.
• Schedule: inrun-db-backup.timer — Mon & Fri 15:00 SGT.
• Local keep: /opt/backups/inrun/<stamp>/, last 12 runs (~100 MB).
• GitHub: private investconfig-spec/sportstlm-db-backups, pushed from
  /opt/backups/github/inrun-db.
• Cost: ~4 s wall, ~150 MB peak RSS.

2. Code snapshots

• /opt/backups/code_backup.py, hourly (code-backup.timer).
• Mirror: /opt/backups/github/code-backups/<module>/ — 17 modules:
  sportstlm, racing, lottery, ufc, snooker, nba, ops, supervision, backup-scripts, tennis, nfl, nhl, mlb, badminton, livenba, inlive2, local-feeds.
• Excludes: data/, *.db, *.sqlite*, .env*, vps-credentials.txt, media,
  node_modules/, .next/, __pycache__/, browser_profile/, {src.
• Credential scan every run (OpenAI/Stripe/GitHub/AWS/Slack/Google keys,
  private-key blocks, Telegram tokens) — aborts the commit on a hit.
• Push target: investconfig-spec/sportstlm-code-backups.

3. Hub DB snapshots (daily)

• /opt/backups/scripts/hub_db_backup.py, daily (hub-db-backup.timer 19:30 UTC).
• Consistent snapshot of every module DB (SQLite VACUUM INTO inside the container,
  or Python sqlite3.backup() read-only for non-node/exited containers), gzipped and
  pushed to sportstlm-db-backups under hubs/<module>/.
• Coverage (as of 2026-11-01): ufc (staging/archive/ufcstats/previews), snooker
  (staging/archive/previews), sportstlm (inlive2 sidecar + aps_staging + logos),
  nba (staging/previews/metrics), tennis (lta + tennis), nfl, nhl, mlb
  (staging), badminton, livenba, lottery, inlive2.
• Retention: 8 snapshots per DB kept locally.

Deliberately not backed up: racing's data/ (daily-regenerated scrape JSON,
covered by the staging "observed-only" policy) and the 0-byte app.db/rsssports.db.

Deploy keys

code_backup.py and the supervisor probe both use the _code key for the code repo.
Probing the code repo with the _backup key returns ERROR: Repository not found —
that means wrong key, not a missing repo.

GitHub PAT alternative

/etc/inrun-backup/github.token (0700 dir / 0600 file), installed via
/opt/backups/accept_credential.py --from-file|--from-url.

Retention / deletion policy

Automated deletion only:
- docker image prune -af --filter until=24h (daily 04:22) — images only
- docker builder prune -f (Wed 04:54) — build cache only
- logrotate — docker/caddy/apt logs only
- systemd-tmpfiles — /tmp 10 d, /var/tmp 30 d
- inrun DB archive rotation — keeps 12 archives

Staging datasets (/opt/sportstlm/data, /opt/racing/data, /opt/lottery/data) are
observed only — no timer/heal may purge them (fallback datasets).