Web Operations. Keeping the Data On Time John Allspaw, Jesse Robbins

Spis treści książki

• Web Operations: Keeping the Data on Time
  • SPECIAL OFFER: Upgrade this ebook with OReilly
  • Foreword
  • Preface
    • How This Book Is Organized
    • Who This Book Is For
    • Conventions Used in This Book
    • Using Code Examples
    • How to Contact Us
    • Safari Books Online
    • Acknowledgments
  • 1. Web Operations: The Career
    • Why Does Web Operations Have It Tough?
      • A Strong Background in Computing
      • Practiced Decisiveness
      • A Calm Disposition
    • From Apprentice to Master
      • Knowledge
      • Tools
      • Experience
        • The organizational challenge of inexperience
        • The concept of "senior operations"
      • Discipline
    • Conclusion
  • 2. How Picnik Uses Cloud Computing: Lessons Learned
    • Where the Cloud Fits (and Why!)
      • Storage
      • Hybrid Computing with EC2
    • Where the Cloud Doesnt Fit (for Picnik)
    • Conclusion
  • 3. Infrastructure and Application Metrics
    • Time Resolution and Retention Concerns
    • Locality of Metrics Collection and Storage
    • Layers of Metrics
      • High-Level Business or Feature-Specific Metrics
      • System- and Service-Level Metrics
    • Providing Context for Anomaly Detection and Alerts
    • Log Lines Are Metrics, Too
    • Correlation with Change Management and Incident Timelines
    • Making Metrics Available to Your Alerting Mechanisms
    • Using Metrics to Guide Load-Feedback Mechanisms
    • A Metrics Collection System, Illustrated: Ganglia
      • Background
      • A Quick Introduction to Ganglia
        • The need to keep collection and aggregation costs low
        • The need to automatically discover new nodes and metrics
        • The need to match network transport with your metrics collection task
        • The need to implicitly prioritize cluster metrics
        • The need to aggregate and organize metrics once they're collected
        • The need to provide convenient interfaces for creating new metrics and pulling out existing metrics for correlation against other data
    • Conclusion
  • 4. Continuous Deployment
    • Small Batches Mean Faster Feedback
    • Small Batches Mean Problems Are Instantly Localized
    • Small Batches Reduce Risk
    • Small Batches Reduce Overhead
    • The Quality Defenders' Lament
      • Why Does It Work?
    • Getting Started
      • Step 1: Continuous Integration Server
      • Step 2: Source Control Commit Check
      • Step 3: Simple Deployment Script
      • Step 4: Real-Time Alerting
      • Step 5: Root-Cause Analysis (Five Whys)
    • Continuous Deployment Is for Mission-Critical Applications
      • Another Release? Do I Have To?
      • The QA Dilemma
    • Conclusion
  • 5. Infrastructure As Code
    • Service-Oriented Architecture
      • Configuration Management
        • Configuration management is policy driven
        • System automation is configuration management policy made into code
        • Configuration management in system administration
      • System Integration
        • Step 1: Break the infrastructure down into reusable, network-accessible services
          • The bootstrapping service.
          • The configuration service.
        • Step 2: Integrate the services together
    • Conclusion
  • 6. Monitoring
    • Story: "The Start of a Journey"
    • Step 1: Understand What You Are Monitoring
    • Step 2: Understand Normal Behavior
    • Step 3: Be Prepared and Learn
    • Conclusion
  • 7. How Complex Systems Fail
    • How Complex Systems Fail
      • (Being a Short Treatise on the Nature of Failure; How Failure Is Evaluated; How Failure Is Attributed to Proximate Cause; and the Resulting New Understanding of Patient Safety)
        • Complex systems are intrinsically hazardous systems
        • Complex systems are heavily and successfully defended against failure
        • Catastrophe requires multiple failuressingle-point failures are not enough
        • Complex systems contain changing mixtures of failures latent within them
        • Complex systems run in degraded mode
        • Catastrophe is always just around the corner
        • Post-accident attribution to a "root cause" is fundamentally wrong
        • Hindsight biases post-accident assessments of human performance
        • Human operators have dual roles: as producers and as defenders against failure
        • All practitioner actions are gambles
        • Actions at the sharp end resolve all ambiguity
        • Human practitioners are the adaptable element of complex systems
        • Human expertise in complex systems is constantly changing
        • Change introduces new forms of failure
        • Views of "cause" limit the effectiveness of defenses against future events
        • Safety is a characteristic of systems and not of their components
        • People continuously create safety
        • Failure-free operations require experience with failure
      • As It Pertains Specifically to Web Operations
        • It will be difficult to tell that the system has failed
        • It will be difficult to tell what has failed
        • Meaningful response will be delayed
        • Communications will be strained and tempers will flare
        • Maintenance will be a major source of new failures
        • Recovery from backup is itself difficult and potentially dangerous
        • Create test procedures that front-line people can use to verify system status
        • Manage operations on a daily basis
        • Control maintenance
        • Assess performance at regular intervals
        • Be a (unique) customer
    • Further Reading
  • 8. Community Management and Web Operations
  • 9. Dealing with Unexpected Traffic Spikes
    • How It All Started
    • Alarms Abound
    • Putting Out the Fire
    • Surviving the Weekend
    • Preparing for the Future
    • CDN to the Rescue
    • Proxy Servers
    • Corralling the Stampede
    • Streamlining the Codebase
    • How Do We Know It Works?
    • The Real Test
    • Lessons Learned
    • Improvements Since Then
  • 10. Dev and Ops Collaboration and Cooperation
    • Deployment
    • Shared, Open Infrastructure
    • Trust
    • On-call Developers
      • Live Debugging Tools
      • Feature Flags
    • Avoiding Blame
    • Conclusion
  • 11. How Your Visitors Feel: User-Facing Metrics
    • Why Collect User-Facing Metrics?
      • Successful Start-ups Learn and Adapt
      • Performance Matters
      • Recent Research Quantifies the Relationship
    • What Makes a Site Slow?
      • Service Discovery
      • Sending the Request
      • Thinking About the Response
      • Delivering the Response
      • Asynchronous Traffic and Refresh
      • Rendering Time
    • Measuring Delay
      • Synthetic Monitoring
        • When to use synthetic monitoring
        • Limitations of synthetic monitoring
        • Configuring synthetic monitoring
      • Real User Monitoring
        • When to use RUM
        • Limitations of RUM
        • Configuring RUM
    • Building an SLA
      • Apdex
    • Visitor Outcomes: Analytics
      • How Marketing Defines Success
      • The Four Kinds of Sites
      • A (Very) Basic Model of Analytics
      • Correlating Performance and Analytics by Time
      • Correlating Performance and Analytics by Visits
    • Other Metrics Marketing Cares About
      • Web Interaction Analytics
      • Voice of the Customer
    • How User Experience Affects Web Ops
      • Many More Stakeholders
      • Monitoring As Part of the Life Cycle, Not Just QA
    • The Future of Web Monitoring
      • Moving from Parts to Users
      • Service-Centric Architectures
      • Clouds and Monitoring
      • APIs and RSS Feeds
        • Delivering an API to others
        • Consuming an API from someone else
      • Rich Internet Applications
      • HTML5: Server-Sent Events and WebSockets
      • Online Communities and the Long Funnel
      • Tying Together Mail and Conversion Loops
      • The Capacity/Cost/Revenue Equation
    • Conclusion
  • 12. Relational Database Strategy and Tactics for the Web
    • Requirements for Web Databases
      • Always On
      • Mostly Transactional Workload
      • Simple Data, Simple Queries
      • Availability Trumps Consistency
      • Rapid Development
      • Online Deployment
      • Built by Developers
    • How Typical Web Databases Grow
      • Single Server
      • Master and Replication Slaves
      • Functional Partitioning
      • Sharding, or Horizontal Partitioning
      • Caching Layer
    • The Yearning for a Cluster
      • The CAP Theorem and ACID Versus BASE
      • State of MySQL Clustering
        • DRBD and Heartbeat
        • Master-Master Replication Manager (MMM)
        • Heartbeat with replication
        • Proxy-based solutions
        • InfiniDB, Galera, Tungsten, and ScaleDB
        • Summary
    • Database Strategy
      • Architecture Requirements
        • Easy wins
      • Safe-Bet Architectures
      • Risky Architectures
        • Sharding
        • Writing to more than one master
        • Multilevel replication
        • Ring replication (beyond two nodes)
        • Reliance on DNS
        • The so-called Entity-Attribute-Value (EAV) design pattern
    • Database Tactics
      • Taking Backups on a Slave
      • Online Schema Changes
      • Monitoring, Graphing, and Instrumentation
      • Analyzing Performance
      • Archiving and Purging Data
    • Conclusion
  • 13. How to Make Failure Beautiful: The Art and Science of Postmortems
    • The Worst Postmortem
    • What Is a Postmortem?
    • When to Conduct a Postmortem
    • Who to Invite to a Postmortem
    • Running a Postmortem
    • Postmortem Follow-Up
    • Conclusion
  • 14. Storage
    • Data Asset Inventory
    • Data Protection
    • Capacity Planning
    • Storage Sizing
    • Operations
    • Conclusion
  • 15. Nonrelational Databases
    • NoSQL Database Overview
      • Pure Key/Value
      • Data Structure
      • Graph
      • Document Oriented
      • Highly Distributed
    • Some Systems in Detail
      • Cassandra
      • HBase
      • Riak
      • CouchDB
      • MongoDB
      • Redis
    • Conclusion
  • 16. Agile Infrastructure
    • Agile Infrastructure
      • But Agile Is Not the Only Thing That Has Evolved
      • Some People Are Born to Web Operations, Some People Have Web Operations Thrust upon Them...
      • Working Software Is the Primary Measure of Progress
      • The Application Is the Infrastructure, the Infrastructure Is the Application
    • So, What's the Problem?
      • Talk Does Not Cook Rice
        • The infrastructure is an application
        • Version control: The foundation of sanity
        • Configuration management and automated deployments
        • Monitoring
        • Dev-test-prod life cycle, continuous integration, and disaster recovery
        • Radiate information
        • Reflective process improvement
        • Incremental changes and refactoring
        • The simplest thing that could work
        • Separation of concerns
        • Technical debt
        • Continuous deployment
        • Pairing
        • Managing flow
    • Communities of Interest and Practice
    • Trading Zones and Apologies
      • What to Do?
    • Conclusion
  • 17. Things That Go Bump in the Night (and How to Sleep Through Them)
    • Definitions
    • How Many 9s?
    • Impact Duration Versus Incident Duration
    • Datacenter Footprint
    • Gradual Failures
    • Trust Nobody
    • Failover Testing
    • Monitoring and History of Patterns
    • Getting a Good Night's Sleep
  • A. Contributors
  • Index
  • About the Authors
  • Colophon
  • SPECIAL OFFER: Upgrade this ebook with OReilly

pokaż cały spis treści

ukryj recenzje