AWS EC2 Skills
Management and orchestration patterns for Elastic Compute Cloud.
🚀 Core Commands
Instance Discovery
# List running instances with Name and Public IP
aws ec2 describe-instances --filters "Name=instance-state-name,Values=running" --query "Reservations[].Instances[].{Name:Tags[?Key==\`Name\`].Value | [0], IP:PublicIpAddress, ID:InstanceId}" --output table
# Find expensive instances (G or P family)
aws ec2 describe-instances --query "Reservations[].Instances[?contains(InstanceType, 'g') || contains(InstanceType, 'p')].[InstanceId, InstanceType]"
Lifecycle Control
# Start/Stop instances by ID
aws ec2 start-instances --instance-ids <id1> <id2>
aws ec2 stop-instances --instance-ids <id>
# Terminate instance (DANGER)
aws ec2 terminate-instances --instance-ids <id>
Network & Security
# Describe security group rules
aws ec2 describe-security-groups --group-ids <sg-id> --query "SecurityGroups[].IpPermissions"
# Add ingress rule (Port 22 from specific IP)
aws ec2 authorize-security-group-ingress --group-id <sg-id> --protocol tcp --port 22 --cidr <your-ip>/32
🧠 Best Practices
- Tag Everything: Use standard tagging (Name, Env, Owner) for billing and discovery.
- Instance Profiles: Use IAM Roles instead of storing hard-coded credentials on instances.
- Spot Instances: Use Spot for stateless workloads (like Remotion rendering) to save up to 90%.
- Security Groups: Default to "Deny All" and only open specific ports for required CIDRs.