github.com/m3db/m3@v1.5.1-0.20231129193456-75a402aa583b/src/cmd/tools/carbon_load/README.md (about) 1 # carbon_load 2 3 `carbon_load` is a tool to generate load on a carbon ingestion port. 4 5 # Usage 6 ``` 7 $ git clone git@github.com:m3db/m3.git 8 $ make carbon_load 9 $ ./bin/clone_fileset -h 10 11 # example usage 12 # ./carbon_load \ 13 -target="0.0.0.0:7204" \ 14 -numWorkers="20" \ 15 -cardinality=1000 \ 16 -name="local.random" \ 17 -qps=1000 \ 18 -duration="30s" 19 ``` 20 21 # Benchmarking Carbon Ingestion 22 23 The easiest way to benchmark carbon ingestion end-to-end is to: 24 25 1. Use `start_m3` to get an entire M3 stack running, including an aggregated namespace and a carbon ingestion port. 26 27 2. Run this tool with the target pointed at the exposed carbon ingestion port on the `start_m3` stack. 28 29 3. Use `pprof` to take CPU and heap profiles of `m3coordinator` while the `carbon_load` tool is running.