⚙️ Backend Engineering · Distributed Systems

Design multi-region data topologies honestly

Residency, replication lag, failover semantics, and write-conflict strategy before the second region exists.

advanced~50 minBackend EngineersAPI DevelopersPlatform Engineers

Steps

1Map residency requirements per dataset BEFORE topology; law constrains design
2Pick write topology: single-primary simplicity vs multi-primary conflict reality
3Model cross-region latency into user journeys region-by-region
4Define failover triggers, RTO/RPO per tier, and split-brain prevention mechanism
5Rehearse regional evacuation with production-scale data annually
6Instrument per-region error/latency separately; averages hide one dead region

Common Pitfalls

▲Multi-primary chosen then conflicts resolved by whoever writes last
▲Failover tested with empty databases only

Commands

Install with skills CLI

$ npx skills add aniruddhaadak80/skills --skill distributed-systems-multi-region-data-topology

Install globally

$ npx skills add aniruddhaadak80/skills --skill distributed-systems-multi-region-data-topology -g

Tags

#multi-region#residency#failover#backend-engineering#distributed-systems

Install this skill

$ npx skills add aniruddhaadak80/skills --skill distributed-systems-multi-region-data-topology -a opencode

SKILL.md

# Design multi-region data topologies honestly

> Residency, replication lag, failover semantics, and write-conflict strategy before the second region exists.

## When to Use This Skill

Residency, replication lag, failover semantics, and write-conflict strategy before the second region exists.

## Steps

1. Map residency requirements per dataset BEFORE topology; law constrains design
2. Pick write topology: single-primary simplicity vs multi-primary conflict reality
3. Model cross-region latency into user journeys region-by-region
4. Define failover triggers, RTO/RPO per tier, and split-brain prevention mechanism
5. Rehearse regional evacuation with production-scale data annually
6. Instrument per-region error/latency separately; averages hide one dead region

## Common Pitfalls

- Multi-primary chosen then conflicts resolved by whoever writes last
- Failover tested with empty databases only

View source on GitHub ↗

Strong vs bounded-staleness vs eventual mapped to business harm scenarios, not fashion.

⚙️ Backend Engineering·~45m

Open →

Adopt event sourcing without the graveyard mistakesadvanced

Command/event separation, upcasting discipline, and projection rebuild strategies that survive year three.