Santekno
Projects/Promotion & Marketing Tools Support Campaign Scenario
ByteDance logoPart of ByteDance · Software Engineer, Promotion Platform Team
Backend / APILive

Promotion & Marketing Tools Support Campaign Scenario

Planned and delivered a complex multi-region campaign scenario in two phases, resolving cross-region routing, timeout configuration, and tag lifecycle integration.

2025-campaign-marketing-tools · main
# E-Commerce / Campaign
$ go -X POST /2025-campaign-marketing-tools/v1/sync
→ 201 Created
{ "status": "ok" }
publish 2025-campaign-marketing-tools.synced
Role
Backend Engineer
Timeline
2025
Team
4
Domain
E-Commerce / Campaign
Stack
8 tech
Status
Live
01

Overview

The Promotion & Marketing Tools Support Campaign Scenario was a complex initiative spanning multiple cases, cross-service dependencies, and multi-region deployment. I owned end-to-end technical planning and execution, designing the tech plan from scratch and splitting it into two clear phases to control risk and keep delivery predictable.

A backend has no screens, so its "UI" is the operational surface: the campaign execution API, multi-region ops, and monitoring dashboards.

API contract — campaign scenario execution
Ops view — multi-region deploy & cross-region routing
Grafana — timeout & latency monitoring
03

Problem & Solution

Problem

The scenario involved many interacting cases and cross-service dependencies deployed across multiple regions, where region-specific configurations and cross-region routing could easily cause regressions or environment leakage. Delivering it safely required careful phasing, strict environment isolation, and tight cross-team coordination.

Solution

I led the full technical design — defining execution flows, mapping cross-service dependencies early, and setting clear boundaries — then aligned stakeholders through a formal tech review. I ensured independent per-region test coverage, validated pre-production environment isolation and data-center alignment to prevent isolated test environment traffic leaking to Production, and collaborated with the tag-management-service team to jointly debug tag association / tag removal flows. Phase 1 shipped on schedule; for Phase 2 I owned the recovery, resolving the QA blockers.

04

Workflow

Three actors interact with the service; the diagrams are written in English to stay consistent across both languages.

flowchart diagram

configure

tag association / removal

cross-region routing

tag lifecycle

Campaign Ops

Campaign Core Service

Tag Management Service

Regional Data Center

Tag Store

Use case interaction
sequence diagram
Timeout Config PlatformTag Management ServiceCampaign Core ServiceCampaign OpsTimeout Config PlatformTag Management ServiceCampaign Core ServiceCampaign Opsrun campaign scenarioapply timeout configtag association / removaltag lifecycle resultexecution status
Sequence flow
05

Architecture

flowchart diagram

metrics

Campaign Console

Campaign Core Service · Go

Timeout Config

Tag Management Service

Data Center · Region A

Data Center · Region B

Data Center · Region C

Prometheus / Grafana

System architecture
Architecture notes

campaign-core-service (Go / RPC) drives the scenario, applying timeout configuration and calling tag-management-service for tag lifecycle operations across data centers in multiple regions. Because tag-management-service runs in another region, traffic is routed cross-region, with tag-association / tag-removal permissions explicitly granted and Prometheus/Grafana monitoring in place.

state diagram

scope A

scope B

timeout / routing fixed

planned

phase1

phase2

stabilized

live

Phased delivery lifecycle
06

Features

Planning
  • Two-phase execution design
  • Cross-service dependency mapping
Multi-region
  • Per-region test coverage
  • Pre-production environment isolation & data-center alignment
  • Cross-region routing
Integration
  • tag-management-service tag association / removal
  • Timeout config & permission setup
07

Tech Stack

language
Go
framework
RPC framework
protocol
RPC
platform
Config / Timeout PlatformTag Management Service
infra
Multi-region
observability
PrometheusGrafana
08

Engineering Challenges

01
Multi-region test isolation and pre-production leakage risk

Each region needed independent, complete test coverage while isolated test environment traffic risked leaking to Production. Fix: enforced data-center alignment and validated pre-production environment isolation, coordinating with QA to avoid region-only regressions.

02
Phase 2 timeout and cross-region routing blockers

Phase 2 hit QA blockers from complex flow dependencies: `campaign-core-service` → `tag-management-service` timeouts and traffic needing to reach another region. Fix: raised timeout config, resolved cross-region routing, and updated tag-association / tag-removal permissions.

03
Cross-team tag lifecycle correctness

Correct tag association / tag removal behavior spanned both campaign and tag-management perspectives. Fix: joint debugging with the `tag-management-service` team to verify the full tag lifecycle.

09

Outcomes

2
delivery phases shipped
3
multiple regions covered
100%
campaign case coverage
0
production incidents
10

Lessons Learned

I gained hands-on experience with multi-region and cross-region execution, learning that correctness in complex campaigns depends as much on environment and infrastructure control as on business logic. Phased execution proved effective at containing blast radius, and early cross-team communication plus joint debugging were key to resolving system-level bottlenecks — timeout, permission, and routing — rather than applying temporary workarounds.

11

Related Santekno Tutorials

Interested in working together?

Order, payment, or high-traffic event pipelines — let's talk architecture and trade-offs.