Jump to: navigation, search

Distributed Monitoring

Revision as of 04:52, 4 September 2017 by Yu-kasuya (talk | contribs) (Created page with "= Distributed Monitoring = Current centralized monitoring architecture is suitable for cloud infrastructure, but in NFV, operator want high ... And Distributed approach is .....")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Distributed Monitoring

Current centralized monitoring architecture is suitable for cloud infrastructure, but in NFV, operator want high ... And Distributed approach is ...

Use Cases

  • Micro Burst Traffic
  • Memory Leak
  • Abnormal behaviour of software/hardware
  • ...

Tools

For distributed monitoring, some open source softwares are useful. ...

collectd

collectd is one of the powerful collecting tools for timebase data... collectd has also several plugins and you can use threshold ,notification and python plugin.

collectd's official page

redis

redis is light in-memory database...

redis's official page

scikit-learn

In this page, scikit-learn is recommended as light weight machine learning library for analytics engine.

scikit-learn's official page

Setup

In computing nodes, controller nods and every nodes that you want to monitor, you can setup following example. In this example, ubuntu16.04 is selected as each node's OS.

  1. set up OpenStack environment using DevStack or manual installation
  2. install collectd, redis and some other python library
  3. get demo code of DMA
# apt install collectd redis python-pip python-dev 

collectd

collectd.conf

 <Plugin "write_redis">
   <Node "dma">
       Host "localhost"
       Port "6379"
       Timeout 1000
   </Node>
 </Plugin>
}

redis

redis.conf

#save 300 10
#save 60 10000