USENIX Courses

Learning More from Complex Systems

Explore complex systems' traits and learn better approaches to incident analysis beyond linear root-cause methods. Gain insights from history, science, and philosophy to enhance understanding of resilience in modern organizations.

• YouTube
• 28 minutes
• On-Demand
• Conference Talk

Evolution of Incident Management at Slack

Explore Slack's evolution in incident management, covering strategies for handling numerous incidents, team-wide capability building, and future directions in maintaining platform reliability.

• YouTube
• 28 minutes
• On-Demand
• Conference Talk

Demystifying Machine Learning in Production - Reasoning about a Large-Scale ML Platform

Explore techniques for managing and improving reliability in large-scale machine learning production systems, focusing on common failure modes, best practices, and practical strategies for SREs.

• YouTube
• 29 minutes
• On-Demand
• Conference Talk

From 15,000 Database Connections to Under 100 - A Tech Debt Tale

Explore a multi-year tech debt resolution journey, focusing on reducing database connections from 15,000 to under 100. Learn strategies for handling scaling challenges in distributed systems.

• YouTube
• 30 minutes
• On-Demand
• Conference Talk

Let's Bring System Dynamics Back to CS

Explore System Dynamics for modeling feedback loops in distributed systems. Learn its history, tools, and applications to prevent outages and improve software architecture design.

• YouTube
• 27 minutes
• On-Demand
• Conference Talk

MySQL and InnoDB Performance for the Rest of Us

Demystifies MySQL and InnoDB performance, covering storage engines, indexes, and query optimization. Provides practical insights for improving database efficiency and understanding SQL operations.

• YouTube
• 34 minutes
• On-Demand
• Conference Talk

SLX - An Extended SLO Framework to Expedite Incident Recovery

Explore an extended SLO framework (SLX) for managing high-availability systems, featuring innovative concepts like SLF and SLD to build knowledge graphs and expedite incident recovery.

• YouTube
• 31 minutes
• On-Demand
• Conference Talk

Optimizing Cost and Performance with Arm64

Explore how Honeycomb.io optimized cost and performance by migrating to arm64 architecture, reducing compute costs by 40% while improving latency, despite challenges in software compatibility and ecosystem complexity.

• YouTube
• 31 minutes
• On-Demand
• Conference Talk

Take Me Down to the Paradise City Where the Metric Is Green and Traces Are Pretty

Explore OpenTelemetry for cloud-native observability, including its APIs and SDKs. Learn how to implement traces and metrics in Java and Go, with practical examples and insights on framework stability.

• YouTube
• 26 minutes
• On-Demand
• Conference Talk

Cache Strategies with Best Practices

Explore efficient cache strategies for data-intensive services, including item management, TTL optimization, and warm-up techniques. Learn to enhance system performance and availability through real-world examples.

• YouTube
• 26 minutes
• On-Demand
• Conference Talk

A Political Scientist's View on Site Reliability

Explore how political science concepts can provide fresh perspectives on site reliability, team dynamics, and simplifying complex production environments in software engineering.

• YouTube
• 30 minutes
• On-Demand
• Conference Talk

Latency Distributions and Micro-Benchmarking to Identify and Characterize Kernel Hotspots

Explore kernel hotspots using latency distributions and micro-benchmarking. Learn techniques to identify and address OS scale limits for improved system performance on bare metal hardware.

• YouTube
• 26 minutes
• On-Demand
• Conference Talk

When Linux Memory Accounting Goes Wrong

Debugging Linux memory accounting issues in cgroups, exploring scenarios where misaccounting can lead to host inaccessibility and application failures despite set limits.

• YouTube
• 27 minutes
• On-Demand
• Conference Talk

Engineering Onboarding

Industry experts discuss challenges, changes, and future trends in engineering onboarding, covering remote work, diversity, training methods, and strategies for improving new hire experiences.

• YouTube
• 50 minutes
• On-Demand
• Conference Talk

Don't Follow Leaders or - All Models Are Wrong - and So Am I

Reflective analysis of SRE book's impact, addressing its strengths, weaknesses, and unresolved challenges in production engineering and reliability practices five years after publication.

• YouTube
• 45 minutes
• On-Demand
• Conference Talk