Separate request, receipt and source times in an operations ledger — what should the operator record?
A late-arriving response can contain older evidence, while a quickly returned page can describe a recent source event. Keep request, receipt and source timestamps separate. The operational ledger should explain processing order without implying that network arrival order establishes the order in which public events actually occurred or became available.
Track latency-related timestamps without replacing source publication time or inferring event order from arrival.
Use request_started_at, response_received_at, source_published_at and locally_committed_at only when actually known. Keep timezone and timestamp precision explicit; a missing service time must not be invented from the interval between two local observations on the operator's machine.
Attach field definitions and actual available timestamps to the artifact manifest. Do not infer exact server processing duration, source discovery delay or population recency from local receipt timing unless additional authoritative evidence supports the precise measurement being claimed for this workflow.Name each clock field
Use request_started_at, response_received_at, source_published_at and locally_committed_at only when actually known. Keep timezone and timestamp precision explicit; a missing service time must not be invented from the interval between two local observations on the operator's machine.
Preserve measured boundaries
Local start and receipt times describe what that caller observed, not the service's internal execution timeline. If clocks may differ across operators, record clock_context and avoid comparing their raw timestamps as though they all share verified synchronization or equal precision.
Test arrival ordering
Create labeled fixtures where a later receipt contains an older source publication time. Verify event-order views use the declared source field and processing-order views use receipt sequence, keeping missing source times unresolved rather than falling back silently to arrival time.
Review apparent anomalies
A negative or implausible local duration can indicate clock interpretation or parsing issues. Quarantine that calculation instead of using an absolute value or swapping timestamps automatically, since a cosmetic fix can hide a real mismatch in how the ledger represented time.
Publish clock semantics
Attach field definitions and actual available timestamps to the artifact manifest. Do not infer exact server processing duration, source discovery delay or population recency from local receipt timing unless additional authoritative evidence supports the precise measurement being claimed for this workflow.
Keep the boundaries clear.
- Track latency-related timestamps without replacing source publication time or inferring event order from arrival.
- Operational record: Name each clock field. Context badges do not certify access.
Start with a read-only check: ask your agent to report actual source coverage before collecting records or exporting an audience.
Primary references
Documentation reviewed 2026-11-15. Product capabilities, client availability and plans can change. Third-party marks identify their products, not an affiliation.