Beyond grep: Practical Logging and Metrics
summary
Argues that concrete plumbing matters not buzzwords and demonstrates a working stack: Sentry (raven client) for deduplicated, context-rich error alerts, metrics as time series with Graphite or InfluxDB for rollups and percentiles not averages, and structured logs produced by structlog as JSON then shipped via runit and logstash-forwarder into ELK (Elasticsearch + Logstash + Kibana). Shows how these pieces reduce noise, expose the 0.01% latency tail, and run reliably at small hoster scale.
AI-written summary · about
chapters (8)
- 0:00Introduction to monitoring
- 1:44Error tracking with Sentry
- 6:57Metrics and time-series data
- 14:32Collecting and storing metrics
- 18:05Centralized logging with ELK
- 22:24Structured logging in Python
- 27:20Logging infrastructure setup
- 29:07Integrating monitoring systems
related talks
35mOpen Source Logging and Monitoring
55mBeginner's guide to observability
27mLogs Are Good, Actually
01h 00mApplication Logging & Logstash
38mLogging as a Painkiller
39mFast log management for your infrastructure
26mWhat every developer should know about logging
29mAre logs a software engineer's best friend? Yes - follow these best practices