On-call done right: how even a developer can help
summary
Demo traces a production failure from Kibana logs through IntelliJ to a single swallowed exception that makes the logs say transactions succeed while customers fail. Uses ELK (Elasticsearch/Logstash/Kibana), distributed tracers (Zipkin/Jaeger), profilers, IDE debugger and exception monitors (Sentry), then advocates adding real-time logs metrics and traces to running apps so developers can ask new questions without redeploying. Promotes continuous observability as a fast, tool-driven way to find and fix that one missing log.
AI-written summary · about
chapters (8)
- 0:00Introduction and speaker intro
- 1:30Meet Jamie the on-call developer
- 8:08What is observability?
- 15:04Tools that don't tell the full story
- 22:26Demo: Debugging a swallowed exception
- 29:06From observability to continuous observability
- 30:09Tips for winning at on-call
- 33:46Closing and Q&A
related talks
38mFrom the Exorcist to Hercule Poirot
45mDeveloper Joy – How great teams get s% t done
50mReading code
50mExpert Software Developers' Approach to Error
58mIntelliJ IDEA Tips and Tricks
27mBeyond developer
18mThe Good, the Bad and the Ugly: The 3 Learnings of an SRE
01h 25mWhat I Wish I Knew before Going On-call