AP3
All topics

Production / connecting product work to production reality

Product Operations

Treat rollout, observability, support, and repair as part of the designed experience.

Why this matters

A feature is not finished when it merges. Product operations is the practice of knowing what shipped, whether it works, how it fails, who can intervene, and how to recover without improvising under pressure.

Concepts to learn

  • Production observability and monitoring
  • Feature flags and staged rollouts
  • Release verification and performance budgets
  • Incident response and repair notes
  • Support tools and internal interfaces
  • Safe-off controls and rollback procedures

Apply It in Casset

Casset is the working environment for this skill—not the curriculum itself.

  • Build a quiet internal operations surface over existing health and Redis telemetry: profile/catalog health, recent failures, and deployment context.
  • Design a platform-wide safe-off or suspension path for a malfunctioning Casset App without duplicating the per-artist enable/disable control already in Studio.
  • Consolidate the fragmented feature-flag vocabulary into a documented registry before adding another switch.
  • Create a production release checklist, run it on one change, and document the first incident or near miss it catches.

Suggested project

Product operations notebook

A small operational surface plus the release, incident, and rollback practices around it.

Open project notebook

Evidence of completion

These are completion conditions, not claimed achievements. Each stays planned until it links to a real artifact or verification.

  • implementationA minimal internal health and recent-failure surfaceplanned
  • productionA completed release verification with links to evidenceplanned
  • incidentAn incident or near-miss write-up with follow-upplanned
  • measurementA documented performance or reliability budgetplanned

Open questions

  • What operational context helps a small team act without turning the product into an enterprise console?
  • Which interventions need a UI, and which should remain a runbook or code-owned control?

Related notes

Books, talks, and useful things

  • book
    Release It!Michael T. Nygard
  • article
    Observability Is for Unknown UnknownsCharity Majors

Reflection journal

Taste includes the quality of the recovery path and the confidence of the person operating the product.