I've recently started to drown my head into automation again - this time from the Cloud. This article appeared on the NJINN Blog!
You must automate? Few things in life tend to be “a must”. Maybe some dreaded events like death and taxes…
… and Automation. Otherwise those life immanent events kick in just too soon.
I would like to point out just three ready-to-automate processes which, today, any IT organization in the world is facing and just needs to get right in the beginning to ease engineers’ and operators’ lives.
Service Requests – no: not any!
That’s no news. Automate your Service-Requests. Boring. When I talk with customers about their experiences in automating their request for service creation or change, mostly what I hear is: “We are special. There’s these two steps that are specific, so we need to script anyway. We stick with scripting.
Well, then … 2 things, though: (1) You are not that different. And (2) Well, do script it then. Powerful automation solutions put their focus into step aggregation, auditing of executions, automation repository, permissions, … and let you do the things, you know best when it comes to the nitty-gritty details (Python, Powershell, Perl, …).
The point is not how you want to automate; that point is, that you automate. Making your users use automated delivery when requesting a service reduces your number of tickets, the occupation of your support engineers, the likelihood of errors and the overall quality of the services you deliver.
Here’s some of those well-known, eternally-been-around examples of IT Service Requests to automate:
- User on- and off-boarding into umpteenth systems
- Virtual machine request, delivery, setup
- On-demand test environments
- Increase mailbox and file storage
- Add Firewall exceptions
- Grant access to applications / change permissions
Would you want to get up at 3 o’clock in the morning, just because log files filled up your disk and now your application is down? Or interrupt whatever you’re working on for the third time, because management needs a report on usage numbers and SLAs right now?
While a full hard drive can easily be avoided by automating log rotation or disk expansion, its necessity only becomes obvious once the incident happens. On the other hand, sending out reports every Monday morning not only avoids those pesky “I need it now” interruptions, but actively highlights the service quality you provide.
IT Housekeeping involves plenty of activities. Here are just a few more examples for must-automate IT maintenance tasks:
- LOG Offload/Rotation
- Synchronize permissions over systems
- Cleanup unused resources
- Daily locked user report
- Push data into analytics / generate dashboard
- Restart on crash (and alert)
Disclaimer: Restart without alert is void. Alert without subsequent action is … well … not quite sensible to say the least (I normally set the alert to automatically create an item on the DevOps team’s board).
Compliance and Audit
No, I won’t go into arguing the need for comprehensive audit data on all actions in an IT environment. Arguably, that’s not arguable anyways.
Since May 25th, 2018, however, a new beast is around in the compliance area: Organizations shiver when only its name is spoken aloud:
Any data owner can hold liable any data controller (and data processor on the controller’s behalf) as to where theirdata is stored, processed, forwarded or used. The three most asked questions by any data owner are:
- Where is my data stored?
- Where did it come from?
- (Please) delete all my data! (“please” usually not included)
So, very simply, in order to sit back and relax, here’s three processes that you must automate regarding GDPR:
- Data entry and change throughout the organization, across all data-stores (CRM, tracking data, custom applications, etc.)
- GDPR compliant access request to all personal data
- GDPR compliant data deletion request
It’s not about survival – maybe!
Whether these examples are “must-automates”, may be negotiable. Less negotiable is the fact, that organizations spend way too much time, effort, frustrating action and negative energy when not doing it. Where does frustration start.
Remember the typical automation evolution path:
- Process knowledge in one’s head
- Recurring Tasks
- Connectivity or credential parametrization
- Repository for collaborative development
- Local execution by multiple people
- Deployment to end-point
I bet, that latest at (4.) the headache begins. Starting at (10.) and working backwards within an enterprise-ready automation solution cures the headache.
Apart from that, what’s your “must-automate-s”?