Google Cloud Profiler

What Google Cloud Profiler is

Google Cloud Profiler helps teams find resource-intensive code in production applications by linking CPU usage and memory allocation information to source code. A profiling agent collects data, and a Google Cloud console interface is used to view and analyze profiles. It supports Go, Java, Node.js, and Python, although profile types vary by language; available types include CPU time, heap, allocated heap, contention, threads, and wall time. Supported environments include Compute Engine, Google Kubernetes Engine, App Engine, Managed Service for Apache Spark, Dataflow, and some environments outside Google Cloud, with language-specific availability. The listed operating systems are Linux with glibc and, for some languages and configurations, Linux with musl. Google says CPU and heap profiling overhead during collection is under 5%, and commonly under 0.5% when averaged across execution time and service replicas. Profile data is retained for 30 days, but profiles can be downloaded for longer-term storage. The service supports VPC Service Controls. It is free to use, though quotas and limits apply.

Who it suits

It is designed for teams analyzing resource consumption and performance in production applications. It may suit teams working in its supported languages and deployment environments.

Limits

The free plan retains profile data for 30 days, and quotas and limits apply. Profiles can be downloaded for longer-term storage.

Getting started

1. Visit the Google Cloud Profiler website.
2. Choose a supported language and deployment environment.
3. Use the profiling agent to collect profiles from the application.
4. View and analyze the collected profiles in the Google Cloud console.

Verdict

Cloud Profiler provides production CPU and memory profiles with source-code attribution and a console for analysis. Check language-specific availability, retention, and quotas against your needs.