Developing Culture To Write Reliable And Performant Services At Scale
summary
Shows how observability is implemented for Python services at scale: JSON-structured logs with module/timestamp/level, per-request trace IDs that cut MTTR, function-level traces and timers instrumented with statsd, and dashboards/alerts surfaced in New Relic or similar APMs. Demonstrates emitting booking events to a columnar store like Apache Cassandra or Redshift for analytics, highlights the four golden signals, and claims handling millions of requests while reducing debugging time via standardized formats, RCAs, and automated alerts.
AI-written summary · about
chapters (22)
- 0:00Introduction
- 1:02About Me
- 1:45Why do we write software
- 4:21Monitoring vs Observability
- 5:55Pillars of Observability
- 6:55Logging
- 9:01Logging Example
- 10:57Trace ID
- 13:02Limitations Of Logging
- 14:12Metrics
- 15:55New Relic
- 18:25StatsD
- 19:46Metrics Limitations
- 20:45Events
- 22:11Events Example
- 22:33Event Schema
- 22:49Ticket Booking
- 23:32Building Culture
- 24:22Right Tools
- 25:28Post Incident Reviews
- 25:59Lead By The Example
- 26:51Conclusion
related talks
33mEffective observability in microservice architectures
16mBuilding Observability into Cloud-Native Applications
30mLogging Rethought 2: The Actions of Frank Taylor Jr.
56mFull-stack observability and application monitoring with
31mObservability driven development
49mImproving Observability of Distributed Systems
30mBest practices for application observability at scale
31mObservability for Developers: How to Get There