package
0.0.0-20221107231338-b44688ae358d
Repository: https://github.com/intelsdi-x/swan.git
Documentation: pkg.go.dev
# Packages
No description provided by the author
# README
snap-plugin-collector-specjbb
Swan uses Snap to collect, process and tag metrics and stores all experiment's data. The following documentation will make sense if you are familiar Snap. You can read more about its plugin model here.
Usage
This is a collector plugin for Snap which parses a SPECjbb standard output file and collects available latency metrics.
The SPECjbb standard output file is a SPECjbb controller's run with its output
piped to a file. When creating a task from the Task Manifest, the SPECjbb collector needs a path to this
file in the stdout_file
configuration field. For example:
{
"version": 1,
"schedule": {
"type": "simple",
"interval": "1s"
},
"workflow": {
"collect": {
"metrics": {
"/intel/swan/specjbb/*/min": {},
"/intel/swan/specjbb/*/p50": {},
"/intel/swan/specjbb/*/p99": {}
},
"config": {
"/intel/swan/specjbb": {
"stdout_file": "/tmp/specjbb.stdout"
}
},
"process": null,
"publish": [
{
"plugin_name": "file",
"plugin_version": 3,
"config": {
"file": "/tmp/metrics.out"
}
}
]
}
}
}
To create a task from the Task Manifest above, run:
snaptel plugin load snap-plugin-collector-specjbb
snaptel plugin load snap-plugin-publisher-file
snaptel task create -t task.json
Following metrics are currently available:
Name | Type | Description | Example value |
---|---|---|---|
/intel/swan/specjbb/*/min | float64 | Minimum read latency (in microseconds) | 300 |
/intel/swan/specjbb/*/max | float64 | Maximum read latency (in microseconds) | 640000 |
/intel/swan/specjbb/*/percentile/50th | float64 | The 50th percentile read latency (in microseconds) | 3100 |
/intel/swan/specjbb/*/percentile/90th | float64 | The 90th percentile read latency (in microseconds) | 21000 |
/intel/swan/specjbb/*/percentile/95th | float64 | The 95th percentile read latency (in microseconds) | 89000 |
/intel/swan/specjbb/*/percentile/99th | float64 | The 99th percentile read latency (in microseconds) | 517000 |