Suggest modification to this talk

Title

Description

By Aaron Pfeifer

The good news: you're quickly signing up new customers, you've scaled your Rails app to a growing cluster of 10+ servers, and the business is really starting to take off. Great! The bad news: Just 30m of failures is starting to be measured in hundreds or even thousands of dollars. Who's going to make sure the lights stay on when your app is starting to fall over? Or worse, what if your app is up, but sign-ups, payments, or some other critical function is broken?

Learn how you can build a robust monitoring infrastructure using the Sensu platform: track business metrics in all of your applications, any system metric on your servers, and do so all with the help of BatsD - a time series data store for real-time needs. We'll also talk about how to look at trending data and how you can integrate Sensu against PagerDuty, RabbitMQ, or any other third-party service. Oh, and of course - everything's written in Ruby, so you can even use your favorite gems!

Help us caption & translate this video!

http://amara.org/v/FGaK/

Date

Summarized using AI?

If this talk's summary was generated by AI, please check this box. A "Summarized using AI" badge will be displayed in the summary tab to indicate that the summary was generated using AI.

Show "Summarized using AI" badge on summary page

Summary

Markdown supported

In his talk at RailsConf 2013, Aaron Pfeifer addresses the critical theme of application monitoring through the lens of two open-source Ruby projects, Sensu and BatsD. As businesses grow and applications scale, monitoring not only becomes essential to maintain uptime but also to protect revenue from potential losses caused by application failures. Pfeifer emphasizes the necessity of developing a robust monitoring infrastructure, and urges engineers to prioritize monitoring their systems and business metrics effectively.

Key Points Discussed:
- **Importance of Monitoring**: Engineers often overlook monitoring while focusing on performance and design; however, it is vital to ensure features work correctly in a live environment.
- **Types of Metrics**: Pfeifer identifies three categories of metrics to monitor:
  - **Business Metrics**: Indicators such as revenue and conversion rates that can signal if there’s an outage affecting key performance indicators.
  - **Application Performance Metrics**: Metrics like response times that provide insight into user experience and performance issues.
  - **System Metrics**: These metrics, including those from services like Memcached and Redis, help diagnose infrastructure problems.
- **Tools for Monitoring**: The talk introduces Sensu as a framework designed for cloud-based applications and BatsD as a real-time time series database that complements Sensu by tracking and storing metrics.
- **Architecture Overview**: Pfeifer explains how Sensu operates using a message bus with RabbitMQ, allowing it to monitor servers and perform defined checks. He covers how to integrate various metrics into existing Rails applications easily.
- **Examples of Checks**: The presentation includes examples of writing checks in Sensu, demonstrating its simplicity and effectiveness in monitoring server health and application performance.
- **Leveraging Handlers**: Handlers process the output of checks and can send alerts via integrations like PagerDuty, ensuring responsible notification management.
- **Integration of Metrics**: Pfeifer underscores integrating metrics tracking into the development workflow, advocating for automatic metric tracking during deployment.
- **The Role of Engineers in Monitoring**: Lastly, he urges a cultural shift where engineers embrace monitoring responsibilities rather than relegating them to operations alone.

By concluding this session, Pfeifer illustrates that effective monitoring solutions are not just necessary but should be seamlessly integrated into the development cycle, allowing teams to catch issues proactively and maintain application reliability. This proactive approach ultimately serves to protect the business and enhance user experience, reaffirming the necessity of monitoring as an integral part of application development.

Cancel