AWS SysOps Administrator Interview Questions¶

Systems Management¶

1. What is AWS Systems Manager (SSM)?¶

A tool for visibility and control. * Run Command: Execute scripts on instances without SSH/RDP. * Patch Manager: Automate OS patching windows. * Session Manager: Secure shell access via browser/CLI (No port 22). * Parameter Store: Manage config data/secrets.

2. How do you resize an active EBS volume?¶

You can increase size without detaching. 1. Modify Volume: Request size increase in Console/CLI. 2. Wait: State moves to "optimizing". 3. Extend FS: OS level. * Linux: growpart (partition) -> resize2fs (ext4) or xfs_growfs (xfs). * Windows: Disk Management -> Extend Volume.

3. Backup Strategy for EC2?¶

AWS Backup or Data Lifecycle Manager (DLM). Create a policy: "Take Snapshot every 24h, Keep for 7 days". AWS Backup adds centralized monitoring and audit support (Compliance).

Monitoring (CloudWatch)¶

4. Difference: CloudWatch Logs vs CloudTrail?¶

CloudWatch Logs: Application logs (Apache, IIS), System logs (Message/Syslog), Lambda execution logs.
CloudTrail: Who did what? (API Calls). "User Alice called StopInstances at 2 PM".

5. How to track memory usage of EC2?¶

Default CloudWatch metrics (CPU, Disk, Net) do not include Memory. You must install the Unified CloudWatch Agent on the OS to push custom memory/disk-used metrics.

6. What is the difference between "Rebooting" and "Stop/Start" an instance?¶

Reboot: OS restart. Instance stays on same host. Public IP (if dynamic) stays same. Ephemeral storage persists.
Stop/Start: Application is moved to a new hardware host. Public IP changes. Ephemeral storage is erased. Helpful to fix "System Status Check" failures.

Troubleshooting¶

7. Instance is "Unreachable". Status Checks explain?¶

System Status: AWS side issue (Power, Network, Hardware). Fix: Stop/Start.
Instance Status: OS side issue (Kernel panic, corrupted FS, exhausted memory). Fix: Varies (Reboot, Rescue instance).

8. Explain the "CPU Credit" concept in T-series instances.¶

T2/T3 instances are "burstable". They earn credits when idle. They spend credits when bursting above baseline CPU. If credits run out, CPU is throttled to baseline (can cause unresponsiveness). Fix: Use T2/T3 Unlimited or upgrade instance type.

Intermediate/Advanced¶

9. How do you recover a root password for Linux EC2?¶

You generally use private keys, not passwords. If using password auth and lost: 1. Stop instance. Detach volume. 2. Attach to rescue VM. 3. Chroot into volume or mount. 4. Run passwd user. 5. Reattach.

10. How to troubleshoot High Latency on ELB?¶

SurgeQueueLength (Classic) / TargetResponseTime (ALB): Is the backend slow?
SpilloverCount: Requests rejected because queue full.
EC2 Metrics: is CPU 100%?
Access Logs: Enable ELB Access logs to S3 to analyze individual request timing.

11. What is Trusted Advisor?¶

An automated tool that scans account for best practices: * Cost: Idle instances, unattached EBS. * Security: Open SG ports, MFA on Root. * Fault Tolerance: Snapshots, Multi-AZ. * Performance: Service Limits.

12. How do you ensure S3 buckets are not accidentally deleted?¶

MFA Delete: Requires MFA token to delete object versions or bucket.
Versioning: Keeps history.
Bucket Policy: Deny s3:DeleteBucket for everyone.

13. How to manage "Service Quotas" (Limits)?¶

Use Service Quotas console. You can view default limits and request increases (e.g., Max vCPUs per region) via the console which opens a support case automatically.

14. Explain "Cost Allocation Tags".¶

Tags (Key-Value) applied to resources. In Billing Console, you "Activiate" these tags. AWS generates a Cost and Usage Report (CUR) breaking down costs by these tags (Cost Center, Project, Owner).

15. What happens if a Spot Instance is interrupted?¶

AWS gives a 2-minute warning (via Instance Metadata & EventBridge). You can configure the interruption behavior: Stop, Hibernate, or Terminate. SysOps role: Ensure script handles the warning (graceful shutdown) or use "Spot Fleet" to diversify instance types.

16. How do you grant a user access to the Billing Console?¶

By default, only Root has access. 1. Root user must enable "IAM User/Role Access to Billing Information" in Account Settings. 2. Attach Billing or AdministratorAccess policy to the IAM User/Role.

17. How to implement Cross-Account Role access?¶

Account A (Resource): Create Role "CrossAccountRole". Trust Policy allows arn:aws:iam::AccountB:root. Attach Permissions.
Account B (User): Create Policy allowing sts:AssumeRole on arn:aws:iam::AccountA:role/CrossAccountRole.

18. Determine who deleted an EC2 instance yesterday.¶

Go to CloudTrail. Filter Event History by EventName = "TerminateInstances" and Time Range. The output shows the UserIdentity (ARN / User Name) and SourceIP.

19. What is "Savings Plans" vs RIs?¶

RI: Commit to specific Instance Type (m5.large) + OS + Region. Less flexible.
Savings Plan: Commit to $ / hour spend (Compute SP). Matches any instance type (m5, c5, t3), any region, Fargate, Lambda. Much more flexible.

20. How to setup a custom metric (e.g., Disk Space)?¶

Install CloudWatch Agent on EC2.
Configure amazon-cloudwatch-agent.json.
Ensure IAM Role attached to EC2 has CloudWatchAgentServerPolicy.
Restart agent. Metric appears in "CWAgent" namespace.