USENIX Courses

The 'Success' in SRE Is Silent

Explores strategies for demonstrating the return on investment of availability work and argues that poor ROI discourse can undermine SRE investment and outcomes.

• YouTube
• 30 minutes
• Self-Paced
• Conference Talk

How the Metrics Backend Works at Datadog

A behind-the-scenes look at Datadog’s metrics backend, tracing how Kafka partitioning and failure handling support reliability at massive scale.

• YouTube
• 31 minutes
• Self-Paced
• Conference Talk

Building and Running a Diversity-Focused Pre-Internship Program for SRE

A case study of Meta’s remote, diversity-focused pre-internship program for introducing college students to SRE careers.

• YouTube
• 28 minutes
• Self-Paced
• Conference Talk

Are We There Yet? - Metrics-Driven Prioritization for Your Reliability Roadmap

A metric-driven framework for SRE leaders to prioritize reliability work, make tradeoffs, and build a data-backed roadmap.

• YouTube
• 29 minutes
• Self-Paced
• Conference Talk

Named Data Networking

Explores Named Data Networking as a data-centric architectural alternative intended to replace TCP, including its security model and research agenda.

• YouTube
• 45 minutes
• Self-Paced
• Conference Talk

Release Engineering Best Practices at Google

A snapshot of Google’s release engineering practices, including hermetic builds, gated operations, branching, automated releases, rollout processes, and configuration management.

• YouTube
• 45 minutes
• Self-Paced
• Conference Talk

Efficiently Backing Up Terabytes of Data with PgBackRest

Learn how PgBackRest efficiently backs up and restores PostgreSQL databases at terabyte scale using parallel transfer, archiving, and incremental strategies.

• YouTube
• 35 minutes
• Self-Paced
• Conference Talk

Building a Billion User Load Balancer

Explore how Facebook globally routes and balances traffic with DNS and software load balancers at billion-user scale.

• YouTube
• 35 minutes
• Self-Paced
• Conference Talk

MySQL Automation at Facebook Scale

A look at how Facebook automated MySQL cluster operations across thousands of servers, from provisioning and promotions to lifecycle maintenance and failure recovery.

• YouTube
• 50 minutes
• Self-Paced
• Conference Talk

SRE in the Small and in the Large

Shows how SRE principles apply from single-person startups to large organizations through monitoring, load testing, state management, and distributed-systems tradeoffs.

• YouTube
• 1 hour 26 minutes
• Self-Paced
• Conference Talk

Network-Based LUKS Volume Decryption with Tang

Demonstrates network-based unlocking of encrypted Linux server volumes at boot with Tang and Clevis while keeping data encrypted at rest.

• YouTube
• 31 minutes
• Self-Paced
• Conference Talk

The Devopsification of Windows Server 2016

Explore how Windows Server 2016 reshapes system configuration, deployment, security, containers, and release pipelines for DevOps workflows.

• YouTube
• 45 minutes
• Self-Paced
• Conference Talk

Building a Billion User Load Balancer

Explore how Facebook globally routes and balances traffic with DNS and software load balancers to improve performance, capacity, and reliability at billion-user scale.

• YouTube
• 45 minutes
• Self-Paced
• Conference Talk

Scaling Shopify's Multi-Tenant Architecture across Multiple Datacenters

A case study of how Shopify evolved a multi-tenant Rails application into a multi-datacenter architecture for scalability, resilience, and disaster recovery.

• YouTube
• 39 minutes
• Self-Paced
• Conference Talk

Doorman - Global Distributed Client Side Rate Limiting

Explore Doorman, Google's distributed client-side rate-limiting system, and how clients coordinate globally to keep shared-resource usage within capacity.

• YouTube
• 26 minutes
• Self-Paced
• Conference Talk