Distributed systems, microservices, containers/schedulers, continuous delivery … we’ve been through one paradigm shift after another when it comes to architecture, but when it comes to observability we’re still using crufty old logging and metrics and dashboards that haven't been innovative since the LAMP stack era. And guess what? These tools completely fall apart past a certain level of complexity. Let’s dig into some of the deep technical reasons why this is happening and talk about some newer approaches to debugging complex systems when every single request into a system must be identifiable and aggregatable (e.g. honeycomb, distributed tracing). Why are events better than metrics? What is cardinality and why does it matter? And what is the difference between monitoring and observability, anyhow? Come find out.
OPEN TALK: Observability and the Glorious Future
- PDTJoin on Hopin
Charity is a co-founder and CTO at Honeycomb.io, a tool for software engineers to understand what happens when their code meets production. She has worked at companies like Facebook, Parse, and Linden Lab, as a systems engineer and engineering manager, but always seems to end up responsible for the databases too. Co-author of O'Reilly's "Database Reliability Engineering". Loves free speech, free software and single malt scotch