Talks
Speakers
Events
Topics
Sign in
Home
Talks
Speakers
Events
Topics
Leaderboard
Use
Analytics
Sign in
Suggest modification to this talk
Title
Description
This talk will dive deep into the internals of one of the fastest growing profiling gems: Pyroscope. Pyroscope is unique because it is actually a ruby gem of another ruby gem written in rust; Pyroscope extends the popular rbspy project as a foundation to not only collect profiling data, but also to tag and analyze that data as well. You can think of Pyroscope as what you would get if rbspy and speedscope had a baby. We’ll start with the internals and end with an example of how two flamegraphs can be used to tell a story about your applications performance.
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 the talk 'A Tale of Two Flamegraphs: Continuous Profiling in Ruby,' Ryan Perry provides an insightful exploration into continuous profiling in Ruby using the Pyroscope gem, which leverages features from rbspy and speedscope for performance analysis. He begins by defining flame graphs, illustrating how they represent the time spent in different parts of an application, ultimately revealing optimization opportunities. Key points discussed include: - **Types of Profilers**: Perry distinguishes between instrumented profilers, which instrument methods to capture execution time, and sampling profilers, which capture stack traces at regular intervals, providing a less accurate but broader view of application behavior. - **Continuous Profiling**: He explains how Pyroscope enhances profiling by continuously collecting data from ephemeral sessions without needing specific commands from the application, allowing retrospective queries akin to monitoring tools like Prometheus. - **Pyroscope Features**: Pyroscope organizes profiling data meaningfully through tagging and metadata, enabling developers to analyze specific functions or controllers. Its user interface is designed to present flame graphs and tabular data, offering insights into CPU usage and resource distribution across different operations. - **Storage and Query Efficiency**: The talk elaborates on challenges related to data storage and querying efficiencies, detailing methods like tree-like compression of stack traces to minimize space requirements, and aggregate data collection within defined time chunks to expedite query responses. - **Real-World Examples**: Perry presents practical scenarios, such as analyzing CPU usage across different routes in a rideshare application to identify performance bottlenecks and optimize resource allocation, and shares an anecdote about refining compression settings to enhance performance based on profiling insights. In conclusion, continuous profiling with Pyroscope provides a powerful means for Ruby developers to understand application performance dynamics, address latency issues proactively, and produce significant performance improvements through data-driven decisions. Perry emphasizes that the insights gained from profiling can lead to substantial operational enhancements in software applications, particularly for high-stakes environments prone to performance fluctuations.
Suggest modifications
Cancel
