1
Metrics
Numeric measurements such as feed age, signal count, order latency, rejection rate, and exposure.
Lesson · Algorithmic and Automated Trading
Observability, Logging, and Auditability
Measure data freshness, decision latency, broker latency, error rate, positions, risk, and end-to-end traces while preserving an immutable decision audit trail.
Core concepts
What to understand
Keep the boundaries, state, and authority explicit.
2
Logs and audit
Structured event records containing identities, versions, reasons, and before/after state.
3
Trace
The causal path from market event through strategy, risk, order, and broker response.
Interactive lab
Trading System Observability Lab
Inject a latency or data incident and toggle metrics, logs, traces, and correlation IDs. Measure detection and diagnosis quality.
Trading System Observability Lab
Deterministic synthetic educational data · no broker or live-market connection
SYSTEM MODEL
Process
Instrument the complete decision path
Telemetry should explain both system health and trading behavior.
Measure
Emit service and domain metrics with clear units and thresholds.
Correlate
Carry event, signal, order, and trace identifiers across components.
Alert
Notify on actionable symptoms and link directly to supporting evidence.
Knowledge check
Check your understanding
Select one answer for each question. Feedback appears immediately, and a wrong choice is answered on its own terms before the rule is restated.
1. Which metric directly protects strategy freshness?
A stale feed can make an otherwise valid rule act on old information.
2. A trace is useful because it:
It shows where time and errors occurred along the end-to-end path.
3. An audit event should include:
Reconstruction requires the exact context of the decision.
4. A trace reads 20, 17, 240, 18 and 24 ms across feed, normalize, strategy, risk and broker. What is the end-to-end time, and which span do you open?
20 + 17 + 240 + 18 + 24 = 319 ms. The strategy span holds 220 ms of the 220 ms excess; the trace tells you where to look before anyone guesses.
Practical rules
- Measure both technical and trading-domain health.
- Use correlation IDs across the entire order path.
- Alert on stale data, unknown orders, risk breaches, and state divergence.
Common mistakes
- Logging only exceptions.
- Using unstructured text without stable IDs.
- Monitoring process uptime while ignoring feed freshness and broker state.
References
Official reading
These sources cover algorithmic controls, market access, broker APIs, system operations, or security.
OpenTelemetry · Signals
Official overview of traces, metrics, and logs used to observe distributed systems.
FINRA · Regulatory Notice 15-09
Official guidance on supervision, testing, controls, documentation, and monitoring for algorithmic trading strategies.