Introduction to Observability
Foundations
Explore the basics of observability with Ray and Anyscale in this foundations course.
About this course
Course Modules
Intro And Setup
Learn the fundamentals of observability—metrics, logs, and traces—and how they help you monitor and debug distributed systems. Then set up local Ray observability by installing Ray, launching Prometheus and Grafana, starting a two-node Ray cluster, and using the Ray Dashboard to verify and explore collected metrics.
Intro to Observability with Ray and Anyscale- Metrics, Logs, Traces, and Events
- Setting Up Your Ray Dashboard Locally
Ray Anyscale Introduction
Learn the core observability concepts in Ray—logs, metrics, and events—and how to use the Ray Dashboard to monitor application and cluster behavior. Then compare Ray’s native tooling with Anyscale’s managed, contextualized observability (persistent metrics, workload context, and post-failure visibility) through an example job that triggers memory pressure and OOM failures.
- Ray Observability With Anyscale
- Anyscale Dashboard Overview
- Anyscale Observability Basics
Ray Anyscale Observability In Detail
Learn how to monitor and debug Ray workloads using Ray and Anyscale observability dashboards, with hands-on examples that show when to use each view. You’ll explore Ray Data pipeline execution status, logs, and metrics (and Anyscale-specific workload dashboards) to identify bottlenecks and operational issues.
- Observability for Different Ray Workloads
- Observability of Ray Data Pipelines
- Observability for Model Serving with Ray Data