Network Automation

 

Network automation remains a popular buzzword which can still be regularly heard during most IT project discussions. When we discuss network automation with our OP5 Monitor customers, even though there are several key drivers, the two that are the most important to them are as follows –

  • To save resources and money by automating the most simple recurring tasks.
  • Improve support for agile business practices, removing the cumbersome and long lead times in traditional IT operations.

Network Automation

There are many network automation tools/solutions available like Heat from the OpenStack project and RedHat’s Ansible in addition to solutions from the major infrastructure vendors such as Cisco, HP, IBM.  Regardless of the choice of systems available, network automation has some basic requirements that will cost you dearly to ignore and are worth baring in mind before you proceed.

Even though there are many things that you can technically integrate and automate, you have to ask yourself honestly are they all relevant?  Yes this appears to be common sense, but you’d be surprised how often this is overlooked.

Another overlooked aspect of network automation we witness frequently ,is taking a bad manual process and automating it in the belief that it will automatically become a good process, in the vast majority of instances this is not the case.


Before you start a Network Automation project

It’s imperative that before you start any type of network automation project you make sure you have –

  • A good understanding of the actual processes today, which resources you will require and what you need the automation to ultimately deliver.
  • A strong governance with the project leader for the network automation project having full control over the decision. Most automation projects involve many different stakeholders with differencing perspectives, so make sure  that the project leader has full power to take the necessary decisions.
  • Keep it simple and progress in steps! Network automation is a skill set and requires experience and knowledge. Successful automation requires a good level of business and product domain knowledge.

Starting your Network Automation Project

These basic steps can collectively save your IT operation a lot of time.

  • Configure event handlers instead of  traditional e-mail alerts. When something is detected in the network monitoring system an email is usually sent out, this should be replaced with a script that attempts to restart the service involved. If the check fails three times in a row, restart the service and (this makes no sense)wait with sending the alarm. This drastically reduces the number of alarms and the problem could be automatically resolved using this approach. We had an interesting conversation with one of our customers recently, they calculated the effort and time spent for each incident and they came to the conclusion that by automating in the above fashion, that they were managing to save on average around $2,000 per day. This one simple automation step was saving the company over $730,000 per year.
  • Use log based checks to replace manual every day checklists, most System Admins still have a checklist they go through every morning to verify that backups have been done, firewalls are ok etc. By using log based checks instead you essentially eliminate the need for manual checklists, if the backup sends “backup not OK” the log monitoring system sends an alarm.


Next Step Choose Your Master Database 

A very common network automation task is to automate the actual provisioning and activation of monitoring for your services. There are many ways, tools and architectural designs to achieve this but they all need some basic components:

  • Select your master database – this master database (DB) can actually be hosted in many different places. The normal place would be The configuration management database (CMDB) however we find that many customers actually have their IT inventory database in many places and formats, the important thing is that the database contains unique info, also known as configuration items (CI) and it’s the single place where your organization executes change management.   master da
    • CI Unique Identifier or Identification Code
    • CI Name or Label (often, both, long names and short names)
    • CI Abbreviations or Acronyms
    • CI Description
    • CI Ownership (organizations and people)
    • CI Importance
    • etc.
  • In most cases theconfiguration management database  (CMDB) or similar systems also provides the service of “auto detect”. There can many parallel system in this part of the solution, quite common is that you have one system optimized for auto detection of network elements and another system for your Servers and applications etc. The important part is that all the collected data from these supporting system actually forward their data to the centralized CMDB, so again – we know we have an updated and correct DB to integrate with OP5 Monitor.
  • Push or Pull? There are two main ways to automate configuration changes, either you have your CMDB to push any changes that occurs to the monitoring system or you set it up to “fetch”, i.e. the monitoring system fetches configuration changes on a pre-set schedule. Both ways have their plus and minus.

Automatic Orchestration

Now when you have all this setup next step could be to actually use the monitoring metrics as data to enable automatic decisions, hence orchestration. A simple example would be that your web server is starting to give to long response times (this is being checked by the monitoring system) at the same time monitor also detects that the web server application is running out of memory.

By combining these two metrics the orchestration system will automatically respond to these “alarms” with an action to auto provision more memory to the particular web server vm and if that did not help you can add more action to the orchestration sequence. In essence you create a workflow script that pretty much mirrors the different steps you would have done in a manual process.

Automation is a real time and resource savior and I encourage you to really exploit this field. We have put a lot effort in making sure you can do a lot of automation in OP5 Monitor through the RestAPI.

 


Open Source Network Monitoring  

OP5 Monitor is the preferred Open Source Networking & Server Monitoring tool for large multi-national companies in over 60 markets, to see OP5 Monitor in action for yourself get started here.

Prefer to get more hands on? Then Download OP5 monitor for free to get first hand experience of how OP5 Monitor can help you to identify and troubleshoot your current issues as fast as possible.