The Performance and Scalability Mindset
4 likes1,823 views
The document discusses the importance of performance and scalability in software applications. It provides an overview of New Relic's application performance monitoring (APM) service, which instruments applications to collect metrics on performance and usage. New Relic's APM monitors over 175,000 application processes globally for over 10,000 customers, collecting billions of metrics per day from web pages and applications written in various programming languages. It also discusses lessons learned in scaling New Relic's own systems to handle the large volume of collected metrics data.
The Performance and Scalability Mindset
- 1. The Performance and Scalability Mindset Brian Doll / @briandoll / New Relic
- 2. Performance is Value Performance is Value
- 5. Optimize for Real World Use
- 6. Martin Fowler’s Technical Debt Quadrant Reckless Prudent “We don’t have time “We must ship now for design” and deal with consequences” Deliberate Inadvertent “Now we know how “What’s Layering?” we should have done it”
- 7. What is your perspective?
- 8. What is your perspective? Speed
- 9. What is your perspective? Speed Cost
- 10. What is your perspective? Speed Cost Operational Efficiency
- 12. Who watches the watchmen?
- 13. SaaS Application Performance Monitoring * APM as a Service * In-app agent instrumentation * Support for 5 Languages (Ruby, Java, PHP, .NET, Python) * 175,000+ app processes monitored globally * 10,000+ customers
- 14. By the numbers... * 20+ Billion application metrics collected every day * 1.7+ Billion web page metrics collected every week * Each "timeslice" metric is about 250 bytes * 100k timeslice records inserted every second * 7 Billion new rows of data every day * Data collection handled by 9 sharded MySQL servers
- 15. New Relic Web UI * Ruby on Rails * nginx * Linux * 2 @ 12 core Intel Nehalem CPUs w/ 48Gb RAM
- 16. Data Collector and Web Beacon Services * Servlets on Jetty * App metrics collector: 180k requests per minute, responding in 3ms * Web metrics beacon service: 180k requests per minute, responding in 0.15ms * Sharded MySQL using the Percona build * 9 @ 24 core Intel Nehalem w/ 48GB RAM, SAS attached RAID 5 * Bare metal (no virtualization)
- 18. 1. Keep it simple
- 19. 2. Less is more
- 20. 3. Trendy != Reliable
- 21. 4. Plan for scale
- 22. s s ode Epi New Ja Relic va y 5. Use the right technology Ngin x Je5y R ub Rails for a given task
- 23. Thanks Brian Doll / @briandoll / New Relic