Notes on ‘Practical Approaches to Implementing Automation’

Advantages of starting at the end - a provisioning automation example:
Request <- Approval <- Placement <- Provision <- Notification
 
Start at last step (notification) and work upstream. Working “backwards” avoids downstream issues, and the requirements naturally reveal themselves instead of needing to be discovered.


Selecting Components for an Automation Framework

Request Workflow:
DO NOT use: Email
Use: Infrastructure portal / End-user portal + ticketing / CRM system
Examples: ServiceNow, Remedy

Automation Software:
DO NOT use: Shell scripts, pet programming language
Use: Configuration management software (Raw APIs as necessary)
Examples: Ansible, Terraform, Puppet, Chef, Infrastructure service managers

Repository (configuration, code, etc.):
DO NOT use: File share
Use: Source control system
Examples: Github/Gitlab, Bitbucket

Environment Inventory and as-builts:
DO NOT use: Excel spreadsheets
Use: Configuration management DB (CMDB)
Examples: ServiceNOW, Zenoss, BMC Atrium

Pro-active operation monitoring:
DO NOT use: Human eyeballs
Use: Enterprise resource monitoring and trending tool
Examples: Cloud Insights, On-Command Insight, Graphite/Graphana

Image: Automation Design Architecture


Comments