What it is: A long-running configuration that ensures a desired number of tasks keep running.

What it’s for:

• Maintain availability (replace failed tasks).
• Auto scale tasks based on load.
• Integrate with load balancers.

Key ideas:

• Service maintains desired count.
• Can attach to an ALB/NLB.
• Supports rolling deployments.

Exam cues:

• “keep N tasks running continuously” → ECS service.

Hard words:

• *desired count* /dɪˈzaɪərd kaʊnt/: số lượng mong muốn
• *continuously* /kənˈtɪnjuəsli/: liên tục