Discuss a project

DL-018 · Case study

Architecture audit of a catalogue-mirroring Amazon repricer

An Amazon reseller's repricer copied large parts of Amazon's catalogue and could not keep up. A fixed-scope audit showed the code was sound and the limit was SP-API itself.

Client
Name withheld
Location
US South
Industry
Amazon reseller
Capabilities
01Audit
01

Challenge

  • Repricing lagged behind price changes on a large mirrored catalogue.
  • Owner needed an independent view before funding more development.
  • Unclear whether to fix the existing system or redesign it.
02

Environment

  • Existing repricing system that mirrors Amazon catalogue and offer data.
  • Frequent SP-API calls for catalogue, pricing and offers.
  • Selling across several Amazon marketplaces.
03

Intervention

  • Fixed-scope review of code, architecture, data flows and API usage.
  • Measured how request volume met SP-API rate limits.
  • Wrote a two-option roadmap: targeted fixes or a redesign.
  • Redesign option rebuilt around Amazon's bulk offline reports.
04

Outcome

  • Clear verdict: code quality was not the problem.
  • Bottleneck located in SP-API throughput, not in the application.
  • Owner got a costed choice between fixing and redesigning.
The repricer's workers mirror Amazon catalogue and offer data through SP-API calls that hit rate limits; the redesign option replaces most calls with bulk offline reports.bottleneckredesign optionSP-APIRate-limitedOFFLINE REPORTSBulk dataAUDITCode + API usageWORKERSMirror + repriceCATALOGUE MIRRORDatabasePRICE UPDATESTo Amazon
  1. 01SP-APIRate-limitedWORKERSMirror + repricebottleneck
  2. 02OFFLINE REPORTSBulk dataWORKERSMirror + repriceredesign option
  3. 03AUDITCode + API usageWORKERSMirror + reprice
  4. 04WORKERSMirror + repriceCATALOGUE MIRRORDatabase
  5. 05CATALOGUE MIRRORDatabasePRICE UPDATESTo Amazon
Fig. 19Where the repricer really waitsThe repricer's workers mirror Amazon catalogue and offer data through SP-API calls that hit rate limits; the redesign option replaces most calls with bulk offline reports.Scale NTS

Results, as published.

2
Roadmap options delivered

Stack and services.

Stack06

  • Amazon SP-API
  • Catalog Items API
  • Product Pricing API
  • Reports API
  • Relational database
  • Scheduled workers

Services used03

Tell us what has to work.

A senior engineer reads every brief. Describe the system, what it runs on today and what happens when it fails.

Discuss a project