Skip to main content

​
Morning startup checklist

1

Confirm services are up (API + dashboard)

If you run interactively:
b1e55ed start --no-browser
If you run under systemd, confirm:
systemctl status b1e55ed-api.service --no-pager || true
systemctl status b1e55ed-brain.service --no-pager || true
2

Health check (cron-safe)

b1e55ed health
b1e55ed health --json
What “healthy” typically looks like at the HTTP layer:
curl -s http://127.0.0.1:5050/api/v1/health | jq
Example response:
{
  "version": "1.0.0-beta.4",
  "uptime_seconds": 12.34,
  "db_size_bytes": 123456
}
GET /api/v1/health is intentionally unauthenticated so it can be used by monitors.
3

Verify brain status

curl -s \
  -H "Authorization: Bearer $B1E55ED_TOKEN" \
  http://127.0.0.1:5050/api/v1/brain/status | jq
You’re looking for:
  • a recent last_cycle_at
  • kill_switch_level: 0
4

Run one cycle (optional but recommended before market open)

b1e55ed brain
# or, to include slower producers:
b1e55ed brain --full
5

Review positions

b1e55ed positions
b1e55ed positions --json | jq
Confirm:
  • what’s open
  • stop loss / take profit (if you use them)
  • any obvious exposure concentration

​
Useful commands (copy/paste)

Health

b1e55ed health --json and GET /api/v1/health

Run the brain

b1e55ed brain --full or POST /api/v1/brain/run

Positions

b1e55ed positions --json or GET /api/v1/positions

​
Monitoring hints

​
1) Watch the event stream (SSE)

The SSE stream is the quickest way to see what the engine is emitting (signals, alerts, etc.). 
curl -N \
  -H "Authorization: Bearer $B1E55ED_TOKEN" \
  "http://127.0.0.1:5050/api/v1/events/stream?domain=alert"
You’ll receive lines like: 
data: {"id":"123","type":"alert.*","ts":"2026-02-20T00:00:00Z","source":"...","payload":{...}}

​
2) Use the dashboard

  • Default: http://127.0.0.1:5051
  • Best used for quick visual sanity checks (recent signals, last cycles, producer status)

​
3) Producer health intuition

If a domain looks “silent”, it’s usually because the producer endpoint env var is missing and the producer is emitting no events (often reported as DEGRADED in health metadata).

​
End-of-day checklist

1

Check kill switch status

b1e55ed kill-switch
If it’s non-zero, investigate before letting automation continue.
2

Export (optional)

b1e55ed export karma --format jsonl --include-chain --output ./exports/karma.jsonl
3

Log review (systemd)

journalctl -u b1e55ed-api.service -n 200 --no-pager || true
journalctl -u b1e55ed-brain.service -n 200 --no-pager || true