Hazardous Defaults: Managing Cardinality and Performance for Your Logging Stack

Explore the challenges of configuring monitoring and observability tools for optimal performance in this 32-minute conference talk from the Cloud Native Computing Foundation (CNCF). Learn about the potential pitfalls of using default configurations in instrumented systems that generate large volumes of data. Dive into the structure of the PLG-stack (Promtail, Loki, and Grafana) and discover how unexamined default settings can lead to unexpected performance issues and data loss. Follow along as the speakers demonstrate thoughtful adjustments to address cardinality impacts and improve overall system efficiency. Gain valuable insights with a step-by-step guide to optimizing your logging stack, ensuring you maximize the potential of your monitoring and observability tools.

Hazardous Defaults: Managing Cardinality and Performance for Your Lo... Derek Cavanaugh & Sara Moore