production-agents
Read this series in order.
Each part builds on the one before it. Reading time, level, and the concepts covered are shown for every part below.
Senior
9 articles
205 minutes
46 concepts
Parts
Articles in this series
- Production Agents Overview - The Loop Is 20% of the Work The demo works before the six missing capabilities show up.
- Idempotency & Safe Retries - The Stripe Pattern for Agents Retries are only safe after the action becomes repeatable.
- State Persistence & Agent Memory - The Complete Domain Agent memory is state management with product consequences.
- Human-in-the-Loop Patterns - When Agents Need Judgment Human judgment belongs at risk boundaries, not everywhere.
- Cost Control & Token Budgets - Preventing $10K Surprises Budgets are the circuit breaker for runaway loops.
- Observability & Silent Failures - Catching What Doesn't Crash Silent failures need semantic telemetry, not just stack traces.
- Durable Execution Frameworks - Don't Reinvent the Wheel Durability is a framework decision before it is agent logic.
- Security & Sandboxing - Containing Agent Risk A capable agent needs a smaller blast radius.
- Testing & Evaluation - Validating Agent Behavior Non-determinism still needs a test harness.