Skip to content
ML Academy

Engineering, production, governance, cost, communication · insight-0300

Give every production model a named owner, a runbook and a support path

Evidence type
practitioner heuristic
Difficulty
easy
Roles
ML engineer
Industries
cross-industry

Summary: situation, action, reason, boundary

  1. Situation

    A model alert fires at night, and the person who built it left the company months ago.

  2. Action

    Register an owning team and on-call rotation per model, write a runbook covering alerts, rollback, fallback and stakeholders, and test it by having someone other than the author follow it.

  3. Reason

    Operational ownership and documented procedures reduce time to recover and dependence on individuals.

  4. Boundary

    Any model whose failure affects customers, money or compliance.

Symptom or risk

Nobody knows what the alert means, how to roll back or whom to inform, so a small issue becomes a long incident.

Exceptions

Short-lived experiments can have lighter ownership if they switch off automatically.

Trade-offs

On-call costs people's time; low-severity models can use business-hours support.

Pitfalls

Runbooks that describe architecture but not what to do.

Lessons and concepts