This application take a nmon file and upload it in a InfluxDB database. It generates also a dashboard to allow data visualization in Grafana. It's working on linux only for now.
Go to my github repository Releases section
You can build the binary from source. You need to have a working GO environment, see golang.org installation instructions. Check GOPATH environment variable to be sure it's defined.
go get -u github.com/adejoux/nmon2influxdb
cd $GOPATH/src/github.com/adejoux/nmon2influxdb
go build
You need a working InfluxDB database to use nmon2influxdb.
Both InfluxDB and Grafana are easy to install on linux. On Windows or Mac OS it's doable but more complicated. Using a Docker container is easier.
You can use my docker-influxdb-grafana docker image :
# docker pull adejoux/docker-influxdb-grafana
To start the container :
# docker run -d -p 3000:3000 -p 8083:8083 -p 8086:8086 --name="nmon_reports" -t adejoux/docker-influxdb-grafana
The git repository of this image is available at : docker_influxdb_grafana
Grafana will be available at url : http://localhost:3000
InfluxDB administration interface will be available at : http://localhost:8083
You will need to know the IP address of your boot2docker VM. You will need to use the options in command line. For example, if my VM has ip 192.168.99.100 :
nmon2influxdb -s 192.168.99.100 import test.nmon test2.nmon
For dashboard, you will need to specify the grafana url too :
nmon2influxdb dashboard file --gurl "http://192.168.99.100:3000" test.nmon
#nmon2influxdb
NAME:
nmon2influxdb - upload NMON stats to InfluxDB database
USAGE:
nmon2influxdb [global options] command [command options] [arguments...]
VERSION:
0.8.0
AUTHOR(S):
Alain Dejoux <adejoux@djouxtech.net>
COMMANDS:
import import nmon files
dashboard generate a dashboard from a nmon file or template
stats generate stats from a InfluxDB metric
list list InfluxDB metrics or measurements
help, h Shows a list of commands or help for one command
GLOBAL OPTIONS:
--server, -s "localhost" InfluxDB server and port
--port, -p "8086" InfluxDB port
--db, -d "nmon_reports" InfluxDB database
--user, -u "root" InfluxDB administrator user
--pass "root" InfluxDB administrator pass
--debug debug mode
--tz, -t timezone
--help, -h show help
--version, -v print the version
The import command is the one used to import NMON files in InfluxDB.
You can skip disks metrics with nd option. Can be interesting when you have a system with a lot of disks and where you are interested only in CPU performance.
nmon2influxdb import -h
NAME:
import - import nmon files
USAGE:
command import [command options] [arguments...]
OPTIONS:
--nodisks, --nd skip disk metrics
--cpus, -c add per cpu metrics
The dashboard command has two subcommands to create dashboards in Grafana.
nmon2influxdb dashboard -h
NAME:
nmon2influxdb dashboard - generate a dashboard from a nmon file or template
USAGE:
nmon2influxdb dashboard command [command options] [arguments...]
COMMANDS:
file generate a dashboard from a nmon file
template generate a dashboard from a TOML template
help, h Shows a list of commands or help for one command
OPTIONS:
--help, -h show help
The dashboard file subcommand will create and upload a dashboard from a NMON file.
It's great as a starting point with Grafana.
NAME:
file - generate a dashboard from a nmon file
USAGE:
command file [command options] [arguments...]
OPTIONS:
--template, -t optional json template file to use
--file, -f generate Grafana dashboard file
--guser "admin" grafana user
--gpassword, --gpass "admin" grafana password
--gurl "http://localhost:3000" grafana url
--datasource "nmon2influxdb" grafana datasource
The dashboard template subcommand will create and upload a dashboard from a TOML template file.
NAME:
template - generate a dashboard from a TOML template
USAGE:
command template [command options] [arguments...]
OPTIONS:
--template, -t optional json template file to use
--file, -f generate Grafana dashboard file
--guser "admin" grafana user
--gpassword, --gpass "admin" grafana password
--gurl "http://localhost:3000" grafana url
--datasource "nmon2influxdb" grafana datasource
Dashboards are not always the best way to analyze data. When you have hundred or thousands disks it's better to have a summary of stats.
nmon2influxdb stats -h
NAME:
stats - generate stats from a InfluxDB metric
USAGE:
command stats [command options] [arguments...]
OPTIONS:
--metric, -m mandatory metric for stats
--statshost, -s host metrics
--from, -f from date
--to, -t to date
--sort "mean" field to perform sort
--limit, -l "0" limit the output
--filter specify a filter on fields
Grafana is really great to build templates but you need to learn it to build advanced dashboard. It's pretty easy but I wanted to provide ready to use templates. JSON is pretty hard to read for humans so I implemented TOML templates.
For example, this template will display informations coming from two differents VIO servers :
title = "templated dual vio"
[templates]
[[templates.template]]
# name of the template variable
name = "vios"
# the default nmon2influxdb database in InfluxDB
datasource = "nmon2influxdb"
# query used to retrieve values in InfluxDB
query = "SHOW TAG VALUES FROM CPU_ALL WITH KEY = host"
# Regular expression used to filter the query result
regex = "vios"
[[row]]
title = "LPAR"
[[row.panel]]
title = "shared processor"
[[row.panel.metric]]
measurement = "LPAR"
# use template variable $vios
hosts = ["$vios"]
fields = ["PhysicalCPU", "virtualCPUs", "entitled"]
[[row.panel.override]]
alias = "$tag_host PhysicalCPU"
stack = true
fill = 1
[row.panel.tooltip]
value_type = "individual"
[[row.panel]]
title = "shared processor pool"
[[row.panel.metric]]
measurement = "LPAR"
# use template variable $vios
hosts = ["$vios"]
fields = ["PoolIdle", "poolCPUs"]
[[row]]
title = "SEA"
[[row.panel]]
title = "SEA WRITE"
stack = true
fill = 1
[[row.panel.metric]]
measurement = "SEA"
# use template variable $vios
hosts = ["$vios"]
# regular expressions are used on fields
fields = ["write-KB"]
[[row.panel]]
title = "SEA READ"
stack = true
fill = 1
[[row.panel.metric]]
measurement = "SEA"
# use template variable $vios
hosts = ["$vios"]
# regular expressions are used on fields
fields = ["read-KB"]
[time]
from = "2015-07-06T13:03:56.000Z"
to = "2015-07-07T14:23:44.000Z"Importing nmon files :
# nmon2influxdb import testsrv_141114_0000.nmon testsrv_141115_0000.nmon
##################################################################################
File testsrv_141114_0000.nmon imported !
##################################################################################
File testsrv_141115_0000.nmon imported !
Each '#' character in the output means 10 000 points was inserted in InfluxDB.
Upload a dashboard to Grafana based on a NMON file :
nmon2influxdb dashboard file testsrv_141114_0000.nmon
Dashboard uploaded to grafana
Generate a dashboard file from the NMON file :
nmon2influxdb dashboard file -f testsrv_141114_0000.nmon
Writing GRAFANA dashboard: testsrv_dashboard
Importing a nmon file without the disk data :
# nmon2influx import -nodisks testsrv_141114_0000.nmon
Writing GRAFANA dashboard: testsrv_dashboard
Generating stats for DISKREADSERV metric :
nmon2influxdb stats -m DISKREADSERV -s lpar1
field| Min| Mean| Median| Max|Points #
hdisk10| 0.40| 2.42| 2.10| 12.90| 1200
hdisk11| 0.60| 2.63| 2.20| 14.00| 1200
hdisk12| 0.50| 2.74| 2.30| 16.30| 1200
hdisk13| 0.00| 0.52| 0.00| 16.30| 1200
hdisk7| 0.00| 0.01| 0.00| 0.80| 1200
hdisk8| 0.00| 0.04| 0.00| 1.00| 1200
hdisk9| 0.50| 2.33| 2.10| 11.30| 1200
Generating stats for DISKREADSERV metric on host lpar1 on a time period :
nmon2influxdb stats -m DISKREADSERV -s lpar1 -f "08:04:05,29-May-2015" -t "15:04:05,29-May-2015"
field| Min| Mean| Median| Max|Points #
hdisk10| 2.20| 3.80| 3.65| 5.60| 132
hdisk11| 2.40| 4.10| 3.90| 5.80| 132
hdisk12| 2.30| 4.39| 4.05| 6.50| 132
hdisk13| 0.00| 0.73| 0.00| 13.90| 132
hdisk7| 0.00| 0.02| 0.00| 0.20| 132
hdisk8| 0.00| 0.10| 0.00| 1.00| 132
hdisk9| 2.10| 3.68| 3.55| 5.20| 132
Generating stats for DISKREADSERV metric on host lpar1 and limit output to the 5 most active disk :
nmon2influxdb stats -m DISKREADSERV -s lpar1 -l 5
field| Min| Mean| Median| Max|Points #
hdisk10| 2.20| 3.80| 3.65| 5.60| 132
hdisk11| 2.40| 4.10| 3.90| 5.80| 132
hdisk12| 2.30| 4.39| 4.05| 6.50| 132
hdisk13| 0.00| 0.73| 0.00| 13.90| 132
hdisk7| 0.00| 0.02| 0.00| 0.20| 132
Listing measurements available for a specific host and containing DISK in the name :
nmon2influxdb list measurement --host feddy --filter DISK
measurements
DISKBSIZE
DISKBUSY
DISKREAD
DISKWRITE
DISKXFER
The code is licensed as GNU AGPLv3. See the LICENSE file for the full license.
Copyright (c) 2014 Alain Dejoux adejoux@djouxtech.net