下载PDF
实例探究 > Fanzter scores spot on performance with New Relic

Fanzter scores spot on performance with New Relic

技术
  • 分析与建模 - 实时分析
  • 应用基础设施与中间件 - API 集成与管理
  • 功能应用 - 远程监控系统
适用行业
  • 零售
  • Software
适用功能
  • 产品研发
  • 销售与市场营销
用例
  • 预测性维护
  • 实时定位系统 (RTLS)
  • 资产跟踪
服务
  • 软件设计与工程服务
  • 系统集成
挑战
Fanzter, a private software company, was under pressure to maintain high performance for its products, introduce new features, and increase application stability. The company's products, which include Coolspotters.com, CoolPapers, Streaks, and Summizer, served over 20 million users in 2011, serving more than 32 million page views each month. High performance for traditional and mobile access was essential to their success. In the early days of Coolspotters, the team analyzed Rails application logs daily to determine the areas with slowest performance. They also tried to correlate the slow query logs from the database with the actions that triggered them. These methods were time consuming and far from foolproof. Fanzter needed a better way to analyze performance, to find and fix performance problems.
关于客户
Founded in 2007 and now headquartered in Charleston, South Carolina, Fanzter is a private software company that creates some of the most popular websites and mobile apps centered around fan expression and content discovery. The small company has a big following with products like Coolspotters.com, CoolPapers, Streaks, and Summizer. Coolspotters.com, the company’s flagship product, connects pop culture and products. Its community of “spotters” creates a wiki-esque catalog of products and fashions used by today’s most popular celebrities. Users of the site can then buy the products they discover directly from the site. Coolspotters is a mashup of celebrity fashion and consumerism. Content from Coolspotters also flows into their popular iOS app, CoolPapers — an app for discovering fun and unique wallpapers for iPhones and iPads. Two additional iOS apps were introduced in 2010 — Streaks, a motivational calendar and goal tracker, and Summizer, a tool for searching and tracking topics and trends on Twitter.
解决方案
Fanzter turned to New Relic, a new startup specializing in production performance monitoring for Ruby on Rails. New Relic provided a simple Rails plug-in which allowed Fanzter to be up and running in short order. As New Relic evolved, they moved the installation process to a Ruby Gem package, allowing not only simple install, but making it very easy to keep the agent up to date. Fanzter uses the New Relic RPM to monitor and manage their Coolspotters application, which powers both the Coolspotters.com website and the API behind the CoolPapers iOS apps. The New Relic Ruby agent runs in more than 30 processes across the stack. Their applications (aggregating web and mobile) deliver more than 32 million page views to over 2 million unique users each month. The vast majority of this traffic is measured by New Relic RPM in at least one dimension. Powering mobile APIs is a big part of what Fanzter’s platform does and New Relic is a serious part of making sure that mobile performance is spot on.
运营影响
  • Using a power tool like New Relic RPM has enabled Fanzter to manage performance real-time, to quickly identify performance issues and just as easily to find and fix root causes.
  • Transaction tracing is particularly useful as it provides a breakdown of the request’s timeline — especially for slow SQL queries or external requests. It eliminates guesswork which improves productivity and enables quick action and remedy for performance optimization.
  • When Fanzter implemented their Holiday Gift Guide for the 2011 season, it included some new views. They quickly discovered that some of these views, which had not been slow in the test environment, were slow once put into production. The problem only occurred in certain categories of the holiday guide when using the full data set. Using Transaction Traces to analyze these requests, the Fanzter team was able to see that a specific SQL query, which had worked well for most categories, was not performing well in a small number of cases. Fanzter adjusted the index usage in the query, implemented a bit more caching, and was able to bring request times back in line with the rest of the site.
数量效益
  • Fanzter served over 20 million users in 2011, serving more than 32 million page views each month.
  • Their applications (aggregating web and mobile) deliver more than 32 million page views to over 2 million unique users each month.

相关案例.

联系我们

欢迎与我们交流!

* Required
* Required
* Required
* Invalid email address
提交此表单,即表示您同意 IoT ONE 可以与您联系并分享洞察和营销信息。
不,谢谢,我不想收到来自 IoT ONE 的任何营销电子邮件。
提交

Thank you for your message!
We will contact you soon.