📄️ Overview
Welcome to Zymtrace! This guide covers all the features and capabilities that help you analyze performance, identify bottlenecks, and optimize your applications using continuous profiling data.
📄️ Efficiency IQ
Identify the top 5 most expensive functions and entities that need your attention now
📄️ Recommendations
AI-generated performance reports that show where each application spends CPU and GPU time, and how to reduce it
📄️ Filtering Data
Slice and dice your data efficiently
📄️ Top Entities
Get a macro-level overview of performance insight
📄️ K8s Resource IQ
Compare Kubernetes CPU and memory requests against real usage and get right-sized values for every container
📄️ GPU Topology
See how GPUs, NVLink, and NUMA nodes are connected on each host, spot degraded links and misplaced processes, and get fixes
📄️ Top Functions
What's the most expensive function across your entire fleet?
📄️ Diff Functions
Compare performance between two distinct states
📄️ Off-CPU Profiling
Find where threads wait on locks, I/O, the network, or a CPU, with off-CPU flamegraphs
📄️ Allocation Profiling
Sample-based allocation profiling for Java workloads (OpenJDK, Zulu, and Zing)
🗃️ Flamegraphs
2 items
📄️ Unit Reference
Definitions of key metrics and units used throughout Zymtrace.