Observability for Emerging Infra - What Got You Here Won't Get You There

Explore the evolving landscape of observability in modern infrastructure through this thought-provoking conference talk. Delve into the limitations of traditional monitoring tools in the face of increasingly complex distributed systems, microservices, and dynamic infrastructure patterns. Discover why conventional metrics, logs, and APM tools fall short in addressing the challenges of today's fluid and transient environments. Learn about the industry-wide shift from monitoring to observability, and understand the importance of focusing on individual request health rather than system-wide metrics. Gain insights into emerging approaches such as event-based monitoring, structured data analysis, and the role of raw request data in troubleshooting. Examine the impact of these changes on team structures, software development practices, and the skills required for effective observability. Challenge traditional notions of dashboards and best practices, and explore innovative solutions for navigating the complexity of modern systems. Prepare to rethink your approach to debugging and monitoring as you confront the "complexity Armageddon" and embrace new paradigms in observability.

Introduction
Monitoring is dead
I hate monitoring
Observability became contested
Complexity is exploding
What is observability
Why we drink
Monitoring is opera
Observability is about
Complexity is rising
How to measure complexity
Lampstack vs Parse
Example problems
Best practices
Examples
PTSD
Unknown Unknowns
Symptoms
Instrumentation
Distributed Tracing
Metrics
Metrics vs Events
Cardinality
Platforms
Structured Data
Dashboards
Dashboard Rot
Why I Hate Dashboards
Every Dashboard is an Artifact of a Past Failure
Dashboards Must Die Once I Am Queen
You Really Want Raw Requests
Every Single Needle Should Be Findable
Right Time Aggregation
Black Swans
Test in Production
Generalist Software Engineers
Excessive Min
WorkLife Balance
Watching It Run
Distributed Systems
What You Win