Execution trace

Validation and efficiency computation are invoked for a particular production snapshot revision. Each attempt is represented by an operation run with an operation key, source revision, input hash, times, status and result reference. Numbered operation steps record the actions inside the run. State history keeps the source’s before and after status with an actor and reason. Together, those records let an analyst answer what was attempted, which revision it touched and what was persisted.

A run marked Failed records a failure reason. Reconciliation required signals that the analyst must compare the operation’s recorded result with the source’s actual state. A Succeeded efficiency run follows a read-back of the updated result fields. The visible state of a run is therefore part of the handoff, not a decorative success message. The analyst should inspect it before asking a manager to rely on the result.

• Run identity. Operation key and input hash distinguish an intended action from a new input.
• Step receipt. Numbered steps preserve input and output references and any external receipt evidence.
• Transition. State history ties a status change to the source revision and actor.

Choose the operation key convention used when an operator retries validation. Agree who investigates Failed and reconciliation required runs, and which receipt evidence is required for an external step. Define how long a run may stay Running before investigation. The analyst and Admin should know which fields to compare on a read-back: the source revision, relevant measured values, result reference and completed step outcomes. Keep run troubleshooting within the authorized organization and source.

Invoke validation for a source with linked facts. Open the operation run and follow its steps in order. Confirm the run points to the intended snapshot and revision. Inspect state history for the recorded transition. Repeat the same action with the same operation key and verify that the existing operation identity explains the response. Then rehearse a failed update: the analyst should see the failure reason and avoid treating the source as analyzed until the persisted fields have been checked. Compare the time and step number on the failed attempt with the earlier successful attempt. This makes a retry understandable to the next shift without treating every request as a new calculation.

• Run exceptions — Failed or reconciliation-required runs with their source revision and reason.
• Production snapshot queue — Sources whose lifecycle state still needs work.

Follow observation validation and efficiency calculation. The data model shows how run, step and source references join.