Interfaces

RunbookWalkOptions

Interface: RunbookWalkOptions

Defined in: src/core/runbook/types.ts:96

The walk policy.

Properties

cap?

readonly optional cap?: number

Defined in: src/core/runbook/types.ts:99

Row cap on the minted walk (default 500). When the full walk does not fit, the CONTROL FLOW survives — see walk.ts for the projection law.


recording?

readonly optional recording?: boolean | RunbookRecordingOptions

Defined in: src/core/runbook/types.ts:138

ALSO file the inner chart's own RECORDING (9.79.0) — { snapshot, events, structure }, the shape observeRecording() mounts — under kind 'recording/run', and put its ref on the spine as result.walk.recording_ref.

true for the defaults, or { label, maxBytes } to set them yourself.

── Off by default, and why that is not timidity ───────────────────────── The walk is a PROJECTION: eight declared columns per row, values off by construction (narrative({ includeValues: false })), so a walk carries sentences about what happened and no payload from it. A recording is the run: the chart's shared state, its whole commit log, and every attached recorder's data. Filing one is a materially bigger promise — it carries whatever the chart WROTE — so it is a thing an operator declares, never a thing a library starts doing to them. Unset, not one extra line runs: no second snapshot is taken, no bytes are measured, no store call is made, and the envelope is byte-identical to 9.78.0.

── What it buys ──────────────────────────────────────────────────────── The row projection cannot be drawn. structure — the chart's build-time graph — is the only route to a drawable flowchart, and no snapshot carries it; a consumer handed 129 rows can only correctly REFUSE to infer the step graph from them. With the recording filed, the lens/explainable-UI flow components mount the runbook's walk as the flowchart it actually ran.

── Redaction, once, for both ─────────────────────────────────────────── The recording's snapshot is read from the REDACTED MIRROR (getSnapshot({ redact: true })), so the redact policy that scrubs the walk scrubs the recording by the same rule at the same moment. One policy, one meaning, both artifacts.

── Best-effort, and the absence is STATED ────────────────────────────── mintWalk's own law: no store, an over-size refusal, or a failed mint costs the REF, never the answer — and walk.recording_note says which, so a reader never has to guess why a ref is missing. With this option unset, the descriptor says nothing about a recording at all.

On this page