← All stories
● Covered by 1 source · 1 reportMedium impact1 neutral

Pinterest and Intel Enhance gProfiler for Internet-Scale Open-Source Profiling

🔄 Updated 11d ago
New to BrevFeed? We gather this story from every outlet covering it into one summary — ranked by real-world impact, not just the latest headline — so you never miss what matters. What is BrevFeed? →

Key points

  • Pinterest and Intel extended gProfiler, an open-source profiler.
  • New features include dynamic profiling and hardware telemetry.
  • gProfiler unifies performance data across multiple languages and layers.
  • It operates at internet scale for large production fleets.

gProfiler Extension for Internet Scale

Pinterest collaborated with Intel to enhance gProfiler, an open-source continuous profiler. This extension aims to provide production performance visibility across Pinterest's infrastructure, which includes hundreds of thousands of instances supporting 640 million monthly active users. The services are written in Java, Python, C++, Go, and Node.js.

Addressing Fragmentation in Profiling

Previously, engineers used separate tools like perf, async-profiler, and py-spy, resulting in fragmented profiles that were difficult to correlate across different languages and machines. The gProfiler extension addresses this by combining system-wide perf data with runtime-specific tools, producing a unified flame graph across processes on a host.

New Capabilities and Integration

The collaboration added dynamic profiling, hardware-level telemetry, wall-time profiling, and performance-counter flame graphs to gProfiler. It also hardened the system for large production environments. gProfiler acts as an orchestration layer, discovering processes, starting appropriate profilers, merging stack traces, and uploading profiles without requiring application code changes. Integration with Intel PerfSpect allows gProfiler to collect hardware-level information such as cache behavior and memory bandwidth.

Operational Impact

The new capabilities allow engineers to trigger profiling remotely via UI or API for specific services, hosts, or clusters. Agents remain dormant, consuming minimal resources (less than 1% CPU and 100 MB memory), and receive profiling instructions through periodic heartbeats. This transforms profiling into an operational capability, removing the need for engineers to gain shell access to individual machines for performance analysis.

✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →

The daily brief

One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.

One email a day. Unsubscribe in one click, any time.

Today's brief

Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.

~4 min · 3 stories · Nov 21

▶ Play today's brief Listen on Spotify

New every morning, and the back catalogue is archived by date.

Reporting from

Pinterest partnered with Intel to extend gProfiler, an open-source continuous profiler, to provide production performance visibility across large infrastructure fleets. The collaboration adds dynamic profiling, hardware-level telemetry, and wall-time profiling, making it easier to correlate performance data across multiple programming languages and machines. This allows engineers to analyze performance across different layers of modern services without manual correlation of separate profiling tools.