Agent Ops & Governance
Run agents in production without losing sleep.
Observability, guardrails, evaluation, cost and usage tracking, and safety for agents once they're live.
10 lessons · about 3.3 hours
Certificate of completion
Earn your Agent Ops & Governance certificate
Verified edition
Mark every lesson in this course complete. You have finished 0 of 10. 10 to go.
See a sample certificate. Anyone can check a certificate with its code at /verify. It records course completion. It is not an exam or an accredited qualification.
The eight pillars of running agents in production, agent SLIs/SLOs, and a maturity ladder.
Traces and spans for agents, what to record, tooling, privacy, alerts and a repeatable debugging routine.
The eval pyramid, datasets from production failures, LLM-judge calibration, CI gates and online evaluation.
Prompt injection, the lethal trifecta, least privilege, policy enforcement in code and sandboxing.
Risk-tiering actions, approval patterns that avoid fatigue, escalation triggers and audit trails.
Usage ledgers, unit economics, budgets and quotas, and honest footprint estimates.
Resilient calls, idempotency, fallbacks, durable queues, model pinning and chaos testing.
Release bundles, eval-gated pipelines, canaries, rollback and the hidden changes that bite.
The agent register, risk tiers, data governance, key frameworks and an incident runbook.
An eight-pillar launch checklist with evidence, blockers, scoring and a 30/60/90 plan.