L11 Monitoring and Feedback Loops

Metadata

• Lecture ID: L11
• Track: D (Closed-Loop, State, and Feedback)
• Tier: 3
• Duration: 25 minutes
• Prerequisites: L10

Learning Objectives

1. Read closed-loop behavior from runtime traces.
2. Distinguish control-loop output vs monitor-event output.
3. Interpret alert events as operational reliability signals.

Core Concept

• Mental model: monitoring flows emit events only when policy/plant behavior violates expectations.
• Key pitfall: logging everything but defining no actionable alert contract.

Live Demo Mapping

• Primary runnable files:
• examples/tutorial/d_closed_loop_state_feedback/07_feedback_intro.py
• examples/tutorial/d_closed_loop_state_feedback/09_execution_monitoring.py

Runnable Commands

Run from repository root:

pixi run python -m examples.tutorial.d_closed_loop_state_feedback.07_feedback_intro --backend multiprocessing --duration 2
pixi run python -m examples.tutorial.d_closed_loop_state_feedback.09_execution_monitoring --backend multiprocessing --duration 2

What To Observe

• Feedback intro converges toward target with decreasing error.
• Monitoring tutorial emits explicit STUCK alerts during injected degradation interval.
• Alert stream is sparse and event-driven by contract.

Failure Drill

• Tighten stuck threshold and observe alert storm behavior.

Exercise

• Required: identify which steps show controller convergence and where monitoring triggers alerts.
• Stretch: propose one additional alert condition with a typed output field.

Evaluation Rubric

• Pass: learner can explain operational difference between state telemetry and alert events.
• Failure signature: learner cannot map an alert to its triggering state condition.

Follow-up

• Next group: L12-L16 (systems/robotics)