Scalable Architecture

Engineering stores, infrastructure insights and everything you need to know about building scalable architecture, straight from the Hacker News community.

Latest Discussions

Technical Deep Dive11d ago

The nuances of scaling and choosing tech stacks

The discussion centers on the fallacy of 'premature optimization' versus the danger of 'wrong scaling factors.' Participants debate the trade-offs between designing for current needs versus massive future scale, noting that over-engineering for massive scale (e.g., billions of RPS) can introduce unnecessary complexity and cost, while under-engineering leads to painful migrations. Key technical benchmarks like WhatsApp's MQTT usage, AWS IAM's throughput, and the distinction between network capacity and transaction complexity are explored.

Join the thread - 35 comments

🔗 Postgres LISTEN/NOTIFY actually scales

Technical Deep Dive12d ago

Database Scaling and Index Optimization

The discussion highlights critical backend scaling strategies: implementing connection pooling proxies (like PgBouncer) to manage high-concurrency API environments, and the necessity of prioritizing descending selectivity in composite indices to prevent inefficient index scans. Actionable advice includes configuring separate pools for reader/writer workloads and being cautious with tenancy keys in SaaS schemas.

Join the thread - 2 comments

🔗 The startup's Postgres survival guide

Technical Deep Dive11d ago

Pragmatic vs. Scalable Engineering

The thread debates the merits of choosing simple, low-overhead architecture (Go/gRPC) versus building for massive scale. Key insight: Match technology to the specific business constraints and known customer limits to avoid premature optimization, though critics warn this approach can accumulate technical debt.

Join the thread - 5 comments

🔗 Postgres LISTEN/NOTIFY actually scales

Experience Report8d ago

Scaling Django Architectures & Migrations

The thread discusses the practical complexities of managing large-scale Django projects. Key insights include: 1) Using `--no-migrations` and in-memory databases with pytest to accelerate slow model testing; 2) The architectural tension between 'Active Record' patterns and managing complex multi-model business logic; 3) The importance of defining strict app boundaries to prevent 'app sprawl'; and 4) Custom utility development for squashing long migration histories.

Join the thread - 2 comments

🔗 Some more things about Django I've been enjoying

Technical Deep Dive11d ago

Database scaling for time-series metrics

The discussion explores strategies for managing high-frequency metric writes. Key takeaways: spreading load via frequent small writes prevents read-time latency, but MongoDB struggles with large aggregated payloads (1000+ keys). Redis is recommended for flat write performance via individual increments, while Postgres and MongoDB serve as viable persistence layers depending on isolation requirements.

Join the thread - 3 comments

🔗 Show HN: Trifle – Open-source analytics that stores answers, not events

Innovation Discussion6h ago

Scaling AI Skills for Teams

Users are proactively solving the 'individual-centric' gap in current AI tools by creating centralized repositories of company-specific skills and plugins. Actionable insight: There is a growing market need for team-based AI orchestration and shared skill libraries to move beyond isolated individual use.

Join the thread - 2 comments

🔗 Agent skills that bring team coding standards to Claude Code and Codex