API testing for a travel-booking aggregator

The client is a UAE travel-booking aggregator that stitches together hotel, flight, and car-rental inventory from dozens of third-party supplier APIs. For an aggregator, a wrong price or a stale availability response is a booking that fails at checkout and a customer lost to a competitor.

Each supplier evolved its API on its own schedule, so integrations broke unpredictably and the causes were hard to pin down.

Third-party suppliers changed their schemas and pricing formats without warning, so integrations broke in production and the aggregator only learned of it when a booking failed. The blast radius of a single supplier change was unpredictable.

Bugs surfaced deep in the data flow (a mismatched schema here, a pricing field there) where they were expensive to trace and reproduce. Testing endpoints in isolation missed the cross-supplier integration failures entirely.

The aggregator needed contract and integration testing that pinned every supplier response to an agreed shape and caught schema and pricing mismatches before they reached a customer.

We established consumer-driven contracts with Pact for each supplier integration, so the aggregator's expectations of hotel, flight, and car-rental responses are pinned explicitly and any drift in a supplier's schema fails verification rather than a booking.

We tested the integration end to end, following a search-to-book journey across suppliers, asserting schema shape, pricing consistency, and error handling at each hop so mismatches surface in the pipeline instead of at checkout.

The suite runs headless through Newman on every build, exercising the REST endpoints and the Pact contracts together, and every failure ships with the exact request, supplier, and expected-versus-actual response so causes are traced in minutes, not days.