Reliability
Observability and safety guardrails
See the alerting, logging, and kill-switch layers that protect users as automation ramps up.
Telemetry stack
- Structured logging (JSON) streams into PostHog for real-time dashboards (Tinybird pipes are optional for external analytics).
- Health probes track data ingestion, analysis queues, and worker uptime.
- Alerting fanout to Slack/Discord when latency spikes or jobs back up.
Risk controls
- Circuit breakers pause scanning or execution on repeated API failures.
- Kill-switches (manual + automated) stop autonomous actions when thresholds breach.
- Pre-trade microstructure checks validate spreads, staleness, and depth before execution.
Human in the loop
Every automated decision surfaces context to operators. Nothing ships trades without explicit approval today.