This workshop is a hands-on deep-dive into the major pillars of observability. Peter Bourgon will explore the defining characteristics of metrics, tracing, and logging, and use that knowledge to instrument an example service, including interfacing with the Fastly stats and logging pipelines. Attendees will leave with specific, actionable intelligence on how to make their own systems more observable and easier to operate at scale.
54. NYC
presents
● RED or USE methods
● Be cognizant of label cardinality
● Be thoughtful about what you instrument
○ Fewer, more considered metrics are generally better
Metrics methodology
Observability workshop
66. NYC
presents
Observability
Summary
Further reading...
● Monitoring and Observability, Cindy Sridharan
● Monitoring isn't Observability, Baron Schwartz
● Metrics, tracing, and logging, Peter Bourgon
● USE and RED methods, Baron Schwartz
● Ch. 6: Monitoring Distributed Systems, Google SRE book
● Instrumentation Best Practices, Prometheus