Over the last few posts I covered the services that keep an AWS account safe and tidy. Today is a recap: what each one does, and how they fit together.
The series at a glance
| Service | What it answers |
|---|---|
| IAM Identity Center | Who can sign in, and to which accounts and roles |
| CloudWatch Alarms + SNS | Is something wrong right now, and who gets told |
| CloudTrail | Who did what, and when |
| AWS Config | What does the setup look like, and does it follow the rules |
| Systems Manager | How do I manage and patch my instances without logging in one by one |
| Secrets Manager and Parameter Store | Where do passwords and settings live, outside the code |
| AWS KMS | Who holds the keys that protect the data |
| AWS Backup | Can I get the data back if it is lost |
How they fit together
- Access: IAM Identity Center gives people the right access, and CloudTrail records what they do.
- Visibility: CloudWatch Alarms tell you when something looks wrong, and SNS sends the message.
- Rules: AWS Config checks resources against rules, so drift gets noticed.
- Operations: Systems Manager lets you run commands and patch at scale.
- Secrets and keys: Secrets Manager and Parameter Store keep values out of the code, and KMS protects them.
- Recovery: AWS Backup keeps recovery points, which can be encrypted with KMS.
Quick questions to ask yourself
- Who can access this account, and is it least privilege?
- If something changes, will I see it in CloudTrail?
- Will an alarm tell me before a user does?
- Are my secrets out of the code, and is the key policy tight?
- Have I tested a restore?
Hands-on: a simple self-check in your own account
- In CloudTrail, confirm a trail or event history shows recent activity.
- In AWS Config, check whether recording is on and look at one rule.
- In CloudWatch, list your alarms and see who gets the SNS message.
- In Secrets Manager or Parameter Store, make sure no secret sits in plain text in your code.
- In KMS, read one key policy.
- In AWS Backup, check that a plan exists and that its resources are assigned.
- Write down one gap and fix it this week.
Common mistakes
- Turning services on once and never looking at them again.
- Alarms that send to an inbox nobody reads.
- Secrets in code or in plain environment values.
- Backups that were never restored.
- Leaving practice resources behind and paying for them.
Key takeaways
- These services work as a set: access, visibility, rules, operations, secrets and recovery.
- Least privilege and logging come first.
- Keep secrets and keys out of the code.
- Test your alarms and your restores.
Tomorrow I'll start a new AWS topic.
Part of #120DaysOfAWS.
