Image By Damir Mijailovic on Pexels
Using AWS Systems Manager and Patch Groups
By Morgan Lucas
This is a post intended for this site, as a way to get a feel of using it consistently. Older posts are here.
This is a way to manage your cloud setup and use patch your AWS infrastructure in a way that meets certain parameters, and minimizes time to update systems en masse (I saved about 25% of time at the end of the project).
What can use AWS Systems Manager?
There are multiple operating systems you can choose from; It's the equivalent of saying "Hey, I have a Linux machine here; So search though all of the Linux options only and see if my particular Instance needs a patch." Or "Here's a group of Windows Servers; Check for patches once a week and let me know."
If you have multiple instances with multiple instances, you can put those into Patch Groups and designate a OS to each, however, an instance can only be in one AWS Patch Group (PG) at a time.
The Challenge In Using AWS Systems Manager
The most challenging part was simply setting things up. I did this in two accounts:
In my root account, with the help of this video on YouTube, I simply put in instances, applied patches easily.
In an IAM account, I had to have multiple permissions, and even then it took about 2 days for the instances to be acknowledged and for me to have patches attached.
How did I solve the problem in the IAM instance in the account with IAM access?
The Fix of AWS Systems Manager
I walked away for 2 days and came back to see the problem was solved, and it wasn't continually kicking me to the Quick Start screen (a screen that would be telling me to add instances to monitor...which I already had) or the AWS Parameters Store. The person I was working with also had no idea. But hey, we got there in the end (somehow).
I know hiring managers don't want to hear this, but fellow tech heads know, sometimes, it just works and we don't know how! Working with others will have us reach a conclusion sooner, and sometimes it's something on the provider's side, and we can call attention to that and encourage it be fixed.
Using AWS Systems Manager
Actually into the service, I can put instances into certain PGs to update them on a schedule. Update windows only run 24 times (a use case example: If I wanted a PG to check for updates once a day, it would only run 24 times). It's a way to keep an eye on things, but I see how someone would be annoyed by it.
Overall, I saved about 25% of update time by implementing this.
There are preset baselines for many of the Instances AWS supports, including Linux, Mac OS, and Windows.