Profile Materialize CPU, memory, and binary size using samply, heaptrack, and nm.
npx skills add https://github.com/MaterializeInc/materialize --skill mz-profile
Please help me install this Agent Skill. Skill: mz-profile Source: https://github.com/MaterializeInc/materialize/tree/main/.agents/skills/mz-profile Command: npx skills add https://github.com/MaterializeInc/materialize --skill mz-profile
Diagnosing slow queries, high CPU usage, and excessive memory consumption in Materialize requires the right profiling tools and invocation flags, which are easy to get wrong. This Skill provides the exact commands and workflows for CPU profiling, memory profiling, and binary size analysis of Materialize processes.
Profile the CPU usage of my running Materialize environmentd process and show me a flame graph.
High-intent search queries and answers about installing and using this skill.
Use samply to profile CPU usage: attach to a running process with samply record -p PID, or record environmentd from startup with bin/environmentd --wrapper "samply record". Samply opens an interactive Firefox Profiler flame graph in the browser when recording finishes.
Run heaptrack on environmentd with bin/environmentd --reset --no-default-features --wrapper "heaptrack --raw", then analyze the output with heaptrack --interpret on the generated .raw.zst file. Heaptrack works but is slow.
Valgrind does not work on Materialize because it is too slow to yield results. Use heaptrack instead, or set MALLOC_CHECK_=3 and MALLOC_PERTURB_=123 with the system allocator to detect some memory corruption issues.
Run nm --demangle=rust --print-size --size-sort --radix=d on the binary to list symbol sizes. Redirect the large output to a file and filter with cut, sort, and uniq to focus on the size of individual symbols.
Pass the wrapper through the orchestrator: bin/environmentd -- --orchestrator-process-wrapper "samply record" for CPU profiling, or use "heaptrack --raw" as the wrapper for memory profiling of clusterd processes.
dprice-devDiagnose Java/Spring Boot performance bottlenecks using metrics and load testing.
OlegHQAnalyze Rust source files to identify performance bottlenecks and refactor suggestions.
LuigiAPCPereiraImplement structured logging, Prometheus metrics, OpenTelemetry tracing, and Grafana dashboards for Go services.
roanbrasilDiagnose production incidents across JVM, Python, Kubernetes, and databases.
