Lodekeeper HEARTBEAT.md — periodic monitoring checklist (GitHub, Discord, CI, PRs)

HEARTBEAT.md

HEARTBEAT.md

⚠️ STEP 1: BACKLOG — DO THIS FIRST, BEFORE ANYTHING ELSE

1. Read BACKLOG.md right now
2. Look for any task that is NOT marked ✅ and is NOT a passive "monitor/watch" item
3. If an actionable task exists → WORK ON IT. Do not proceed to monitoring.
   • Set it to "in progress" in BACKLOG.md immediately
   • Update dashboard: ~/lodekeeper-dash/scripts/update-status.sh working "<task>"
   • Do the work. Reply with what you did. NOT HEARTBEAT_OK.
4. If the only remaining items are passive monitoring (awaiting review, watching threads) → proceed to Step 2
5. Add any new tasks discovered from notifications/Discord

The failure mode to avoid: Running all the monitoring checks below, seeing "nothing new", and replying HEARTBEAT_OK while an actionable task sits untouched in the backlog. Monitoring is Step 2. Backlog work is Step 1.

Keep dashboard status fresh

• Working on a task → ~/lodekeeper-dash/scripts/update-status.sh working "<task>"
• Genuinely idle (only monitoring) → leave it, auto-idle is correct
• Never set "working" for heartbeat polling itself

STEP 2: Monitoring (only after confirming no actionable backlog tasks)

• Check GitHub notifications (unread OR updated since last read):

  gh api notifications?participating=true --jq '.[] | select(.unread or (.updated_at > .last_read_at)) | {id, reason, title: .subject.title, type: .subject.type}'
  

• If any need attention: review and respond to comments, then mark as done
• Discord @mentions arrive instantly (no polling needed), but do a search backup check:

  message action=search channel=discord query="lodekeeper" guildId=593655374469660673 limit=5
  

Monitor my open PRs CI status

Check CI on all open lodekeeper PRs. If any fail, investigate immediately:

for pr in $(gh pr list --repo ChainSafe/lodestar --author lodekeeper --state open --json number --jq '.[].number'); do
  sha=$(gh api repos/ChainSafe/lodestar/pulls/$pr --jq '.head.sha')
  fails=$(gh api "repos/ChainSafe/lodestar/commits/$sha/check-runs" --jq '[.check_runs[] | select(.conclusion == "failure")] | length')
  if [ "$fails" -gt 0 ]; then echo "PR #$pr has $fails failures"; fi
done

Monitor my Discord threads

Check for new messages in threads I created or where I'm mentioned. Track thread IDs in memory/discord-threads.json.

Monitor unstable branch CI (CONTINUOUS TASK from Nico)

Check latest CI runs on the unstable branch. If any failed, investigate and open a PR to fix.

gh run list --repo ChainSafe/lodestar --branch unstable --limit 10 --json databaseId,name,conclusion,status,createdAt --jq '.[] | select(.conclusion == "failure") | {id: .databaseId, name, created: .createdAt}'

• If failures found: check logs with gh run view <id> --repo ChainSafe/lodestar --log-failed
• Investigate root cause, open a fix PR
• Track investigated failures in memory/unstable-ci-tracker.json to avoid re-investigating

🧹 Periodic cleanup (every ~6 hours)

• Remove completed/merged items from this file and BACKLOG.md
• Move done tasks to the "Completed" section in BACKLOG.md
• Clean up stale daily notes older than 7 days (archive key info to MEMORY.md first)
• Keep this file lean — only active/relevant items