AWSome Day Follow-Up (Part 2 of 2)
A couple of people asked me about the takeaways that *I* think people should have from today's session. Here they are.
- If you ssh or use rdp to update your production environment, your automation is broken. If you think about this, it is easy to see why. If you have automation, and don't fix it, the next time you run it, you'll step on your changes. It is much better to put the update in the pipeline and move them appropriately.
- Always on is the enemy of the cloud. Remember, it's "pay as you go." Why pay for something that you're not using. Granted, there's something to be said for buying in bulk. However, AWS does this for you so that you can save money.
- Tag everything! Inside AWS, you cannot hide your workloads. However, if you want to be able to gather accurate reporting details, you need to tag everything you use with meaningful tags. My favorite tag that I've seen "in the wild" is freshness date. If your instance is past this date, it can be "retired."
- Never use the root account for production work. If it is compromised, there's nothing that can be done to stop it. Instead, create an IAM account with full privileges and use it.
- Amateurs backup their data, professionals restore. Yes, you might have the best disaster recovery (DR) plan on the planet. If you never test it and it doesn't work, your plan is worthless. It will be an RBE. (A Resume Building Event.)