Coordination Health Review¶

A standing review — instantiates Distributed Coordination Architecture

A standing review that watches interface- and outcome-level health signals and re-tunes the coordination architecture before degradation hardens into failure.

Coordination Health Review is the recurring review that treats the coordination architecture itself as a system with a health state, and watches for its decay. It tracks signals that measure the interfaces and the joint result — rejected handoffs, stale shared state, contradictory commitments, wait time, rework, unowned exceptions, overload, externalized cost — and, on a set cadence, decides whether the mode, interfaces, or escalation paths still fit, reconfiguring them (with migration and rollback) before drift becomes breakdown. Its defining discipline is what it refuses to count: meeting attendance and message volume are activity, not health, and both can rise precisely as coordination gets worse. It is the continuous, in-flight counterpart to an after-action review — it runs during standing operation and changes the live architecture, rather than closing out a finished episode.

Example¶

An airline runs a monthly Coordination Health Review over its network operations — the standing coordination among gate crews, fueling, catering, crew scheduling, and the ATC liaison desk. The signal set is deliberately interface-level: the share of aircraft turnarounds with a late or rejected handoff, the average gate-hold minutes attributable to coordination (weather stripped out), the count of contradictory commitments (a crew rostered in two places), and the number of exceptions that went unowned.

The trend that surfaces is a steady climb in rejected fuel handoffs after a new fueling vendor was onboarded. Because the review's cadence includes a reconfiguration decision — not just a chart — it acts: tighten the fuel-handoff interface contract and add a readiness checkpoint before pushback, with a rollback if the checkpoint adds more delay than it removes. Notably, "number of coordination calls held" went up over the same period; a review that watched that activity metric would have concluded coordination was intensifying, exactly backwards.

How it works¶

The signal set measures interfaces and outcomes, not activity. Every signal maps to a coordination failure the parties can act on; attendance and message counts are explicitly excluded as activity theater.
It runs on a standing cadence, not on episode closure — weekly, monthly, quarterly — so decay is caught while it is still cheap to fix.
It attributes. It separates coordination-caused degradation from exogenous shocks, so a bad month isn't misread as a broken architecture (or vice versa).
It carries reconfiguration authority. Each review can hold, adjust, or reconfigure the mode/interfaces/escalation — with migration and rollback — rather than merely reporting a number up the chain.

Tuning parameters¶

Cadence frequency — weekly firefighting versus quarterly drift-watching. Faster catches decay sooner but risks reacting to noise; slower is calmer but lets rot set.
Signal-set composition — which interface and outcome signals, and how many leading versus lagging. Leading signals warn earlier but are noisier.
Reconfiguration threshold — how bad a signal must get before the architecture is changed, trading stability against responsiveness.
Attribution rigor — how hard the review works to separate coordination failure from exogenous causes before acting.
Rollback readiness — how reversible each reconfiguration is kept, which sets how boldly the review can experiment.

When it helps, and when it misleads¶

Its strength is catching coordination decay early and tying the measurement to a structural change — it is a review with authority to re-tune, not a dashboard. Its refusal to reward activity metrics is what lets it see deterioration that a "how many syncs did we hold" view would miss.

Its failure modes are the mirror image of its discipline. Watch vanity or activity signals and the review reports health rising as coordination collapses. Strip the review of reconfiguration authority and it becomes a decorative dashboard that measures but cannot act. Over-react to noise and it churns the architecture on every wobble. And it is easily misused by cherry-picking the green signals to report coordination "healthy" upward. The discipline that keeps it honest is to measure interfaces and joint outcomes rather than activity, and to give the review real authority to reconfigure with rollback.^[1]

How it implements the components¶

coordination_health_signal_set — it defines and watches the interface- and outcome-level signals (rejected handoffs, stale state, contradictory commitments, wait, rework, unowned exceptions, overload, externalized cost) while excluding activity measures.
adaptation_and_reconfiguration_cadence — its standing cadence is where the architecture is reviewed for continued fit and reconfigured, with migration and rollback, before degradation hardens.

It does not close out ended episodes or extract transferable lessons — that is the After-Action Coordination Review; it does not hold the commitments whose contradictions it detects — that store is the Commitment and Dependency Register; and it does not decide who is *allowed to make a reconfiguration — that authority lives in the Coordination Decision Rights and Autonomy Matrix.*

Instantiates: Distributed Coordination Architecture — the standing monitoring-and-adaptation layer that keeps the architecture fit as conditions change.
Consumes: Commitment and Dependency Register — the source of the contradictory-commitment and stale-commitment signals it tracks.
Sibling mechanisms: After-Action Coordination Review · Commitment and Dependency Register · Coordination Decision Rights and Autonomy Matrix · Dependency and Interaction Map · Distributed Planning and Reconciliation Session · Event-Driven Coordination Channel · Exception and Escalation Protocol · Interface Control Document or Service Contract · Joint Operating Agreement · Liaison and Integrator Role · Shared Coordination Board · Synchronization Checkpoint

Notes¶

A health review without authority to reconfigure is just a dashboard. The mechanism only earns its place when the cadence can actually change the mode, interface, or escalation — measurement that cannot act is where coordination decay goes to be admired rather than fixed.

References¶

[1] Goodhart's Law — named for the economist Charles Goodhart and often paraphrased "when a measure becomes a target, it ceases to be a good measure." It is why this review guards against activity signals: the moment "number of coordination meetings" becomes the health metric, it stops measuring coordination and starts measuring meetings, which multiply as things fall apart. ↩