Node.js Q&A
summary
Pino is recommended for high-throughput logging while Log4js gives a nicer API and Winston is fine for everyday use; emit JSON to stdout + Fluentd → Elasticsearch or Splunk for aggregation. Wrap a tiny custom logger as an indirection layer rather than building a framework. Heroku is the fastest path when no DevOps team exists but costs and tool limits grow; AWS gives managed services and compliance at higher cost compared to DigitalOcean. Latency stacks across 5–10 downstream calls so prefer a modular monolith or event-driven design; Apollo subscriptions fall back to websockets and docs can be sparse.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:30Which logger library to use and why
- 2:30Platform agnostic logging
- 3:45Heroku vs Heroic
- 6:00Dedicated Server
- 7:05Microservices
- 9:40Debugging
- 11:55Error levels
- 13:15Custom loggers
- 15:00Latency between microservices
- 16:20Security latency
- 17:50Rest vs Graphql
- 21:10GraphQL documentation
- 29:10Outro
related talks
17mLogging, Metrics, and Tracing with Node.js
37mThe Cost of Logging
02h 24mHow to Node.js - Part 2
22mMulti-tiered Node.js architectures
46mNodeJS Scalibility Tips
16mNode.js in the Cloud
30mNodeJS: Patterns and Tools to Ship to Production with Confidence
24mArchitect for Scale - My Learnings in Node.js (A case study)