Skip to main content
Back to the villageDONGGU’S FIELD NOTESONESELL
The workshop · selected work
PLACE
(주)더커머스
ROLE
Backend
STARTED
2021-03
★ Case 03 / 5Forward Deployed Engineer Donghyun Kang

Onesell Multi-Marketplace SaaS

Held up an environment of 10,000 sellers pushing to 100+ marketplaces, safely, on an async workflow.

(주)더커머스Backend2021.03 ~ 2022.05
01.Discover

The problem was not "send to many marketplaces" — it was "send safely, fast, and traceably."

It looked simple. A seller creates a product, lists it to 100 marketplaces at once, and inventory and orders sync both ways. Once I looked inside, each marketplace had its own API shape, synchronous calls froze the seller's screen on the spot, and when something failed, someone had to dig by hand for which marketplace dropped what.

The problem was not "send to many marketplaces" — it was "send to many marketplaces safely, fast, and traceably." So the work was not to write one good adapter; it was to stand up the full async workflow from zero: request → queue → process → result.

Field notes

Note 01

Every marketplace API was different

Response shape, auth, error codes, retry policy — a hundred different variants per marketplace.

Note 02

Synchronous processing was blocking the seller UX

One click on "list" turned into 100 API calls that the seller had to sit through.

Note 03

Failures never reached a human hand

Which marketplace dropped what either never reached an operator, or reached them late.

02.Build with AI

Built a seat for the operator between the queue and the workers.

Decision 01

Unify the per-marketplace adapter interface

BeforeEach new marketplace scattered call code across the codebase
AfterA single adapter contract — adding a new marketplace lands in one place

tradeoffUpfront abstraction cost vs. long-term maintenance cost

Decision 02

SQS + Lambda async

Seller requests land on SQS immediately and return a response fast. Lambda workers process per-marketplace queues in parallel in the background. The seller's screen is no longer tied to marketplace API response time.

Decision 03

Idempotent Lambdas

An idempotency key keeps the result consistent when the same message arrives multiple times. SQS auto-retry becomes safe, and even when an operator manually replays a message, no duplicate processing happens.

Decision 04

Auto-retry + DLQ + Slack

Delivery failures get 3 SQS auto-retries; once that ceiling is crossed, the message is isolated in the DLQ. When something lands in the DLQ, an #oncall-onsel Slack alert fires immediately — the time from failure to a human seeing it drops to minutes.

Onesell multi-market sync workflow

Onesell · Multi-market sync workflowExample flow
Event flow
Async processing workflow on SQS + LambdaSeller requests are pushed onto SQS, processed asynchronously by Lambda workers, and results returned. Failures auto-retry; exceeding the limit isolates them in a DLQ and alerts via Slack.Create product requestSeller → 100+ marketsSQS queueSeparate per marketAuto-retry x3λ Lambda workerParallel instances · idempotentSuccessInventory & orders syncedMarkets ↔ SellerRetry limit exceededDLQ isolationOperator reviewSlack instant alert#oncall-onsel

Design decisions

  • 01Unified adapter interface per market. Adding a market touches one place in the code.
  • 02Seller requests go straight into the queue; the response returns fast and processing runs async.
  • 03Lambda workers are idempotent. Receiving the same message multiple times yields the same result.
  • 04Send failures auto-retry via SQS; once over the limit they're isolated in a DLQ.
  • 05When DLQ piles up, Slack alerts immediately and the operator reprocesses.

  • Adapter interface
  • Event fan-out
  • Auto-retry policy
  • DLQ isolation
  • Slack alert wiring

Alert triggers → Slack

  • A message lands in the DLQ#oncall-onsel
  • Retry rate exceeds 1%#oncall-onsel
  • Market sync delay exceeds 5 minutes#onsel-ops
03.Prove with metrics

Only the results you can read off an instrument.

Scale

10,000+ sellers

100+ marketplaces in parallel

Stability

Sync → async

Instant seller response, processing in the background

Extensibility

New marketplace

Adapter added in one place

Reliability

DLQ → Slack

Failures caught within minutes

What this case is really about
“A backend engineer who holds stability and traceability together at scale.”
Discover

Read the per-marketplace API differences and the UX cost of synchronous processing as a single pattern.

Build

SQS + Lambda for async, idempotent Lambdas, one unified adapter interface.

Prove

Running for 10,000+ sellers; DLQ-to-Slack catches failures within minutes.

Linked principlesEvidence-DrivenOwnership