Skip to main content

Log & Audit

Log & Audit answers “what happened on this operation?” Status, duration, connection, identifiers, and — when captured — request/response evidence.

Open: app.bundleport.com/connect/transactions/hotels (connect.read). Categories: hotels, flights, ground, packages, transfers.

This screen is the audited grain, not full search volume. API Performance is full volume.

What is always recorded​

OperationIn Log & Audit
Quote, book, cancelAlways
Booking detail, booking list, metadataAlways (metadata has no dedicated operation filter; it shows under Other / unfiltered)
Search / availabilitySampled for high-throughput API-key traffic; always for interactive console traffic

Search modes​

  • By criteria — time range, operation (search, quote, book, cancel, booking_detail, booking_list), status (ok / error), error code or description, connection, duration buckets (over 1s / 3s / 5s / 10s).
  • By ID — transaction_id, booking_id, search_id, quote_id, correlation_id, client_reference, provider_reference.

Context chips can carry booking_id, search_id, quote_id, error_type, error_category. error_category is the same console bucket as API Performance; it is not on Hotels JSON.

The traffic chart and table show time, operation, result, duration, connection, and error text. Export metadata as CSV (10,000 row cap; narrow filters if the export is truncated).

Detail and evidence​

Opening a row shows identifiers, a related funnel timeline when IDs were forwarded, and an evidence package (request/response) when audit bodies were stored.

Pass settings.requestId on the API and keep X-Request-ID when you open a support ticket. See Observability.

Search sampling can hide the exact API-key call you need. From Log & Audit, Audit Search can turn on temporary full capture for a connection when your account can manage Connect. /connect/audit-sampling redirects here with ?audit=1.

Do not treat sampled search as “the API dropped the request.” Check API Performance for full-volume error rates first.