Notification Cadence Guide

Notification Cadence Guide

Turn a cadence_request into a delivery rule for recurring operations updates. The rule should reduce repetitive messages while preserving time-sensitive items.

Cadence decision

Identify the recipients' timezone, desired delivery window, update volume, and the field that identifies duplicate events. Choose a single end-of-day digest when the request favors consolidated reporting. Keep an explicit urgent-item exception only when the request defines what qualifies as urgent.

Grouping and duplicate handling

Group updates by the operational subject named in the request, such as service, site, or owner. Use a stable event or case identifier as the deduplication key; do not deduplicate on free-form message text alone. State the cutoff used to move late items into the next digest.

Output policy

Return cadence_mode as a short string containing the cadence, local delivery hour, timezone, grouping field, deduplication key, and any urgent exception. This output schedules a later composer but sends no notification itself.

Example

For a Shanghai operations team requesting one daily summary, the policy can specify an 18:00 Asia/Shanghai digest grouped by service and deduplicated by incident ID, with updates after the cutoff carried into the next business day.

Interface reference

Input field: cadence_request. Operations update, team notification, or communications handoff.

Accepted value: object.

Output field: cadence_mode; the returned value is a string.

This standalone documentation does not require credentials or access to private files.