23 posts tagged with "database"

Supabase is a Backend-as-a-Service (BaaS) platform that is almost entirely open source, which provides flexibility for developers. While their hosted service is what most people use, you have the freedom to self-host the entire stack. This open-source nature is a key differentiator when comparing it to other solutions like a custom stack with Hasura.

As a Django app grows, managing database queries effectively is crucial for maintainability, performance, and testability. Storing them haphazardly is a recipe for disaster. The best place to store your queries is on custom Model Managers and QuerySets. This approach keeps your logic organized, reusable, and closely tied to the data it operates on. Here are the main variants for storing your queries, from the least recommended to the best practice.

In high-throughput backend services and microservices, data serialization is frequently the primary CPU bottleneck. Converting incoming HTTP payloads or MessagePack buffers into validated Python objects-and marshaling domain entities back into JSON-can consume a significant portion of request processing time.

While Pydantic remains the dominant choice across the Python ecosystem, my experience building backend APIs and data pipelines has led me to evaluate alternative frameworks when hitting latency and memory constraints.

This article provides an engineering analysis of Python data serialization frameworks, evaluating their underlying compilation engines, memory characteristics, developer ergonomics, and suitability for performance-critical systems.