Profiling Kubernetes with cluster, node and pod level filtering using gProfiler
2 by NoamSali | 0 comments on Hacker News.
Granulate has officially released the much-awaited Filters Feature to its gProfiler. In case you need some context, gProfiler(https://ift.tt/3goEoKU) is a free, open-source tool that allows you to run low-overhead, continuous profiling in production to identify bottlenecks, improve code quality, and optimize performance while saving on compute costs. Some of the gProfiler benefits: - Seamless installation with no code changes - Immediate visibility into production code (<5 min.) - Minimal performance overhead and less than 1% utilization penalty - Native support for Java, Go, Python, Scala, Clojure, PHP , and Kotlin With this new feature, gProfiler users can now filter services based on Container name, Hostname, or K8s object. This means that you can view a flame graph of what you want to investigate from a specific container host or pod. To get started, all you have to do is follow these simple steps: - Create your gProfiler account (https://profiler.granulate.io/) - Add a new service using our installation wizard using a Daemonset installation - Inspect your profiling data and start optimizing You can also play around and see how it works in our demo application(https://demo-profiler.granulate.io/). Click on the filters icon and see how the graph filters accordingly. BTW - We’ve also added support for profiling PHP using phpspy, in addition to our existing support for Java, Python and Go, and more.