jmeter性能压测可视化部署(windows)

2019-03-11  本文已影响0人  眉音语

1.安装环境:windows

2.所装软件如下:

influxdb1.4.2:是一个时间序列数据库,用于处理海量写入与负载查询

grafana5.4.2:是一个跨平台的开源的度量分析和可视化工具,可以通过将采集的数据查询然后可视化的展示,并及时通知

telegraf1.9.1:是收集和报告指标和数据的代理,这里主要用于监控服务器的各项指标

chronograf1.7.2:可选,用来操作influxdb的工具,不安装的话,可以通过influxdb自带的工具进行操作)

3、安装步骤

influxdb

①解压后,双击运行influxd.exe

②双击运行influx.exe

通过命令创建数据库:

创建jmeter数据库(用于存储jmeter工具执行的结果数据),命令:create databases jmeter

查看数据库命令:show databases

grafana

①解压后,运行bin目录下grafana-server.exe

②在浏览器输入http://localhost:3000

③默认登录账号:admin;密码:admin

④创建数据源:

新建数据源 选择influxDB 填写数据连接信息

⑤导入模板(模板可去官网下载,json格式)

导入模板 选择新建的数据源jmeter 数据配置

⑥运行jmeter,查看数据是否变化,若有数据,则成功。

在jmeter中--添加--监控器--Backend Listener

telegraf

①讲文件夹放在对应的服务器上

②在influx新建数据库telegraf

③修改配置

# Configuration for sending metrics to InfluxDB

[[outputs.influxdb]]

  ## The full HTTP or UDP URL for your InfluxDB instance.

  ##

  ## Multiple URLs can be specified for a single cluster, only ONE of the

  ## urls will be written to each interval.

  # urls = ["unix:///var/run/influxdb.sock"]

  # urls = ["udp://127.0.0.1:8089"]

  urls = ["http://localhost:8086"]    //此地址为influxdb的ip地址

如果想改变抓取数据频率,则修改如下:

# Configuration for telegraf agent

[agent]

  ## Default data collection interval for all inputs

  interval = "2s"  //默认为10

  ## Rounds collection interval to 'interval'

  ## ie, if interval="10s" then always collect on :00, :10, :20, etc.

  round_interval = true

  ## Telegraf will send metrics to outputs in batches of at most

  ## metric_batch_size metrics.

  ## This controls the size of writes that Telegraf sends to output plugins.

  metric_batch_size = 1000

  ## For failed writes, telegraf will cache metric_buffer_limit metrics for each

  ## output, and will flush this buffer on a successful write. Oldest metrics

  ## are dropped first when this buffer fills.

  ## This buffer only fills when writes fail to output plugin(s).

  metric_buffer_limit = 10000

  ## Collection jitter is used to jitter the collection by a random amount.

  ## Each plugin will sleep for a random time within jitter before collecting.

  ## This can be used to avoid many plugins querying things like sysfs at the

  ## same time, which can have a measurable effect on the system.

  collection_jitter = "0s"

  ## Default flushing interval for all outputs. Maximum flush_interval will be

  ## flush_interval + flush_jitter

  flush_interval = "2s"  //默认为10

  ## Jitter the flush interval by a random amount. This is primarily to avoid

  ## large write spikes for users running a large number of telegraf instances.

  ## ie, a jitter of 5s and interval 10s means flushes will happen every 10-15s

  flush_jitter = "0s"

注:如果想修改jmeter抓取频率则修改jmeter的配置文件:

Summariser - Generate Summary Results - configuration (mainly applies to non-GUI mode)

#---------------------------------------------------------------------------

#

# Comment the following property to disable the default non-GUI summariser

# [or change the value to rename it]

# (applies to non-GUI mode only)

summariser.name=summary

#

# interval between summaries (in seconds) default 30 seconds

summariser.interval=1  //改这里可以改变刷新频率

#

# Write messages to log file

#summariser.log=true

在服务器上,telegraf文件夹下,运行命令:telegraf -config telegraf.conf

上一篇下一篇

猜你喜欢

热点阅读