Topic 04 / 08 · 08 articles

Data

About this topic

130 words

Data modelling decisions are the hardest ones to reverse. A document shape that is convenient for the first screen you build can force expensive fan-out queries forever after, and adding a second database solves one class of problem while creating two new ones.

The articles in this topic look at those decisions in detail: modelling Firestore collections for predictable, indexed reads; when a graph database like Neo4j genuinely earns its place next to a document store; how to keep two stores in sync with events and idempotent upserts; connection pooling for small services; and what happens operationally when a managed database decides to go to sleep.

Each piece ends with a clear rule of thumb for when the approach is worth it — and when a simpler option will do.

01Start here

02All articles in Data