Lessons Learned from Running Prometheus

Explore key insights and practical knowledge gained from operating Prometheus in this comprehensive conference talk by Red Hat experts Lili Cosic and Frederic Branczyk. Delve into the fundamentals of Prometheus, including its server architecture, metadata handling, and alerting capabilities. Discover why Prometheus has gained popularity and examine its simplicity as a single binary with strict rules. Gain a deeper understanding of Prometheus' inner workings, including time series databases, queries, memory usage, and series compression techniques. Learn about resource management, the pull model, retention policies, and target discovery. Analyze scalability considerations and predictability in Prometheus deployments. Explore Prometheus as a platform, its integration with Kubernetes, and how to leverage the broader ecosystem. Conclude with valuable insights and a Q&A session to enhance your Prometheus expertise.

Intro
Prometheus Overview
Prometheus Server
Metadata
Alerting
Prometheus Website
What makes Prometheus simple
Prometheus is a single binary
Prometheus has very strict rules
Why Prometheus is so popular
Prometheus running example
Timeseries database
Queries
Memory Usage
Series vs Samples
Series Compression
Gorilla Compression
Understanding Memory Usage
Memory Map to Disk
Resource Usage
Pull Model
Retention
Target Discovery
Network Routing
Theoretical example
Scale
Predictability
Prometheus as a Platform
Prometheus Kubernetes Integration
Coop Prometheus
Leverage the ecosystem
Conclusion
Questions