Minimum, maximum, and desired capacity in AWS - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is the minimum capacity in an Auto Scaling group?

The minimum capacity is the smallest number of instances that the Auto Scaling group will keep running at all times, even if demand is low.

beginner

Define maximum capacity in the context of Auto Scaling.

Maximum capacity is the highest number of instances the Auto Scaling group can scale out to, preventing it from launching more than this number.

beginner

What does desired capacity mean in an Auto Scaling group?

Desired capacity is the target number of instances the Auto Scaling group tries to maintain, balancing between minimum and maximum capacity.

intermediate

Why is it important to set a minimum capacity in Auto Scaling?

Setting a minimum capacity ensures your application always has enough instances running to handle basic traffic and stay available.

intermediate

How do minimum, maximum, and desired capacity work together?

Minimum sets the floor, maximum sets the ceiling, and desired capacity is the current target number of instances between those limits.

What happens if the desired capacity is set below the minimum capacity in an Auto Scaling group?

A The group will scale up to the maximum capacity.

B The group will maintain the minimum capacity instead.

C The group will scale down to the desired capacity anyway.

D The group will stop all instances.

The Auto Scaling group will not go below the minimum capacity, so it keeps at least that many instances running.

If you want to limit costs but still handle traffic spikes, which capacity setting should you adjust?

A None of the above

B Minimum capacity

C Desired capacity

D Maximum capacity

Maximum capacity limits how many instances can run, controlling costs during traffic spikes.

Which capacity value does Auto Scaling try to maintain under normal conditions?

A Minimum capacity

B Maximum capacity

C Desired capacity

D Zero

Desired capacity is the target number of instances Auto Scaling tries to keep running.

What is the risk of setting minimum capacity too high?

A Higher costs due to more running instances

B Slower scaling

C Instances may stop unexpectedly

D No risk at all

A high minimum capacity means more instances always run, increasing costs even when demand is low.

If traffic suddenly drops, what will Auto Scaling do if desired capacity is above minimum capacity?

A Scale down to minimum capacity

B Scale up to maximum capacity

C Keep current instances running

D Terminate all instances

Auto Scaling will reduce instances but not below the minimum capacity.

Explain in your own words how minimum, maximum, and desired capacity settings affect an Auto Scaling group.

Think of minimum as a safety net, maximum as a limit, and desired as the current goal.

Describe a real-life example where adjusting minimum, maximum, and desired capacity would help manage an application’s performance and cost.

Imagine a store that needs more workers during busy times but fewer when quiet.

Practice

1. What does the MinSize setting control in an AWS Auto Scaling group?

easy

A. The exact number of servers to run at all times

B. The largest number of servers allowed to run

C. The smallest number of servers that must always be running

D. The number of servers to add during scaling

Solution

1. Step 1: Understand MinSize meaning

   MinSize defines the minimum number of servers that the Auto Scaling group keeps running to handle load.

2. Step 2: Compare with other settings

   MaxSize is the maximum allowed, DesiredCapacity is the target number, so MinSize is the smallest number always running.

3. Final Answer:

   The smallest number of servers that must always be running -> Option C

4. Quick Check:

   MinSize = smallest server count [OK]

Hint: MinSize = minimum servers always running [OK]

Common Mistakes:

• Confusing MinSize with DesiredCapacity
• Thinking MinSize is maximum limit
• Assuming MinSize is the exact number to run

2. Which of the following is the correct way to set the desired capacity to 5 in an AWS Auto Scaling group configuration snippet?

easy

A. DesiredCapacity => 5

B. "DesiredCapacity" = 5

C. DesiredCapacity: "5"

D. "DesiredCapacity": 5

Solution

1. Step 1: Identify correct JSON syntax

   In JSON, keys and string values use double quotes and key-value pairs use colon (:).

2. Step 2: Check each option

   "DesiredCapacity": 5 uses correct JSON syntax: key in quotes, colon, numeric value without quotes.

3. Final Answer:

   "DesiredCapacity": 5 -> Option D

4. Quick Check:

   JSON key-value uses colon and quotes [OK]

Hint: JSON uses colon and quotes for keys [OK]

Common Mistakes:

• Using equals sign instead of colon
• Putting numbers in quotes unnecessarily
• Using arrow symbols instead of colon

3. Given an Auto Scaling group with MinSize=2, MaxSize=6, and DesiredCapacity=4, what happens if the current number of running servers is 5 and one server fails?

medium

A. The group does nothing because the current number of running servers equals DesiredCapacity

B. The group launches a new server to maintain DesiredCapacity of 4

C. The group terminates another server to reach DesiredCapacity

D. The group launches two new servers to reach MaxSize

Solution

1. Step 1: Understand DesiredCapacity role

   DesiredCapacity is the target number of servers to keep running, here 4.

2. Step 2: Analyze failure impact

   One server fails, so running servers drop from 5 to 4, which matches DesiredCapacity.

3. Step 3: Check if scaling needed

   Since running servers dropped below DesiredCapacity, the group launches a new server to maintain DesiredCapacity.

4. Final Answer:

   The group launches a new server to maintain DesiredCapacity of 4 -> Option B

5. Quick Check:

   Running servers < DesiredCapacity triggers scaling [OK]

Hint: Scaling triggers only if running servers < DesiredCapacity [OK]

Common Mistakes:

• Assuming scaling happens immediately on failure
• Confusing MinSize with DesiredCapacity
• Thinking MaxSize limits minimum servers

4. You configured an Auto Scaling group with MinSize=3, MaxSize=5, and DesiredCapacity=6. What is the problem?

medium

A. DesiredCapacity cannot be greater than MaxSize

B. MinSize cannot be less than DesiredCapacity

C. MaxSize cannot be less than MinSize

D. No problem, this configuration is valid

Solution

1. Step 1: Check relationship between capacities

   DesiredCapacity must be between MinSize and MaxSize inclusive.

2. Step 2: Compare values

   DesiredCapacity is 6, but MaxSize is 5, which is less than DesiredCapacity.

3. Step 3: Identify error

   This violates the rule that DesiredCapacity cannot exceed MaxSize.

4. Final Answer:

   DesiredCapacity cannot be greater than MaxSize -> Option A

5. Quick Check:

   DesiredCapacity ≤ MaxSize [OK]

Hint: DesiredCapacity must be ≤ MaxSize [OK]

Common Mistakes:

• Setting DesiredCapacity higher than MaxSize
• Confusing MinSize and MaxSize roles
• Assuming any values are allowed

5. You want an Auto Scaling group to always run at least 2 servers, never more than 8, and normally keep 5 running. If traffic suddenly spikes, which setting controls the maximum servers it can scale up to, and what should it be set to?

hard

A. MaxSize controls maximum servers; set to 8

B. DesiredCapacity controls maximum servers; set to 8

C. MinSize controls maximum servers; set to 5

D. DesiredCapacity controls maximum servers; set to 2

Solution

1. Step 1: Identify which setting limits maximum servers

   MaxSize sets the upper limit of servers allowed in the group.

2. Step 2: Match requirement to setting

   To never exceed 8 servers, MaxSize must be set to 8.

3. Step 3: Understand DesiredCapacity and MinSize roles

   DesiredCapacity is the target number (5), MinSize is the minimum (2), neither limits maximum scaling.

4. Final Answer:

   MaxSize controls maximum servers; set to 8 -> Option A

5. Quick Check:

   MaxSize = max servers allowed [OK]

Hint: MaxSize limits max servers; DesiredCapacity is target [OK]

Common Mistakes:

• Confusing DesiredCapacity with MaxSize
• Setting MinSize as maximum limit
• Not setting MaxSize high enough for spikes