Why Memory requests and limits in Kubernetes? - Purpose & Use Cases

The Big Idea

What if one app could silently crash all your others just by using too much memory?

The Scenario

Imagine running many apps on a shared computer without telling it how much memory each app needs. Sometimes one app uses too much memory, making others slow or crash.

The Problem

Manually guessing memory needs is slow and risky. If you give too little memory, apps crash. Too much memory wastes resources. Without limits, one app can hog all memory, breaking others.

The Solution

Memory requests and limits let you tell Kubernetes exactly how much memory each app needs and the max it can use. This keeps apps stable and the system balanced automatically.

Before vs After

✗ Before

Run app without memory settings
kubectl run myapp --image=myimage

✓ After

Set memory request and limit
kubectl run myapp --image=myimage --requests='memory=200Mi' --limits='memory=500Mi'

What It Enables

This lets Kubernetes manage resources smartly, keeping apps healthy and the system running smoothly even with many apps.

Real Life Example

A company runs many microservices on the same cluster. Setting memory requests and limits prevents one service from crashing others during traffic spikes.

Key Takeaways

Manual memory management causes crashes and slowdowns.

Memory requests reserve needed memory; limits cap max usage.

Using them keeps apps stable and resources balanced automatically.

Practice

1. What does the memory request specify in a Kubernetes pod configuration?

easy

A. The total memory available on the node

B. The maximum amount of memory the container can use

C. The amount of memory the container has currently used

D. The minimum amount of memory guaranteed for the container

Solution

1. Step 1: Understand memory request meaning

   Memory request is the minimum memory Kubernetes guarantees to a container so it can run without being killed due to lack of memory.

2. Step 2: Differentiate from memory limit

   Memory limit is the max memory allowed, not the minimum guaranteed.

3. Final Answer:

   The minimum amount of memory guaranteed for the container -> Option D

4. Quick Check:

   Memory request = minimum guaranteed memory [OK]

Hint: Requests = minimum guaranteed memory, limits = max allowed [OK]

Common Mistakes:

• Confusing memory request with memory limit
• Thinking request is the max memory
• Assuming request is current usage

2. Which of the following is the correct way to specify memory requests and limits in a Kubernetes pod YAML?

easy

A. resources:
requests:
memory: "256Mi"
limits:
memory: "512Mi"

B. resources:
memory:
request: 256
limit: 512

C. memory:
requests: 256Mi
limits: 512Mi

D. resources:
requests:
mem: 256Mi
limits:
mem: 512Mi

Solution

1. Step 1: Recall correct YAML structure for resources

   In Kubernetes, memory requests and limits are under resources.requests.memory and resources.limits.memory with values as strings including units like "Mi".

2. Step 2: Check each option for syntax correctness

   resources: requests: memory: "256Mi" limits: memory: "512Mi" matches the correct syntax. Options A, B, and C use wrong keys or structure.

3. Final Answer:

   resources: requests: memory: "256Mi" limits: memory: "512Mi" -> Option A

4. Quick Check:

   Correct YAML keys and units = resources: requests: memory: "256Mi" limits: memory: "512Mi" [OK]

Hint: Use resources.requests.memory and resources.limits.memory keys [OK]

Common Mistakes:

• Using wrong keys like 'mem' or 'memory' outside resources
• Omitting quotes or units for memory values
• Swapping 'request' and 'requests'

3. Given this pod spec snippet:

resources:
  requests:
    memory: "128Mi"
  limits:
    memory: "256Mi"

What happens if the container tries to use 300Mi of memory?

medium

A. The container is allowed to use 300Mi without issues

B. The container is throttled but not killed

C. The container is killed by Kubernetes for exceeding the memory limit

D. The container's memory usage is capped at 256Mi but continues running

Solution

1. Step 1: Understand memory limit enforcement

   Memory limit is the maximum memory a container can use. If it exceeds this, Kubernetes kills the container.

2. Step 2: Apply to given values

   The limit is 256Mi. Using 300Mi exceeds this limit, so the container will be terminated.

3. Final Answer:

   The container is killed by Kubernetes for exceeding the memory limit -> Option C

4. Quick Check:

   Memory usage > limit = container killed [OK]

Hint: Memory over limit kills container, not throttled [OK]

Common Mistakes:

• Thinking container is throttled instead of killed
• Confusing requests with limits
• Assuming container can burst beyond limit safely

4. You set memory requests to 512Mi and limits to 256Mi in your pod YAML. What is the problem?

medium

A. Memory limits cannot be lower than 256Mi

B. Memory requests cannot be higher than memory limits

C. This configuration is valid and will work fine

D. Requests and limits must be equal

Solution

1. Step 1: Recall relationship between requests and limits

   Memory requests must be less than or equal to memory limits. Requests are minimum guaranteed, limits are max allowed.

2. Step 2: Analyze given values

   Requests = 512Mi, Limits = 256Mi. Requests are higher than limits, which is invalid and causes an error.

3. Final Answer:

   Memory requests cannot be higher than memory limits -> Option B

4. Quick Check:

   Requests ≤ Limits rule violated [OK]

Hint: Requests must be ≤ limits, else config error [OK]

Common Mistakes:

• Setting requests higher than limits
• Assuming limits can be lower than requests
• Thinking requests and limits can be any values

5. You want to ensure your pod always has at least 200Mi memory but never uses more than 400Mi. How should you set memory requests and limits in the pod spec?

hard

A. requests.memory: "200Mi" and limits.memory: "400Mi"

B. requests.memory: "400Mi" and limits.memory: "200Mi"

C. requests.memory: "200Mi" only, no limits

D. limits.memory: "400Mi" only, no requests

Solution

1. Step 1: Identify minimum guaranteed memory

   Minimum memory guaranteed is set by requests.memory, so set it to 200Mi.

2. Step 2: Identify maximum allowed memory

   Maximum memory allowed is set by limits.memory, so set it to 400Mi.

3. Step 3: Verify correct order

   Requests must be ≤ limits, so 200Mi ≤ 400Mi is valid.

4. Final Answer:

   requests.memory: "200Mi" and limits.memory: "400Mi" -> Option A

5. Quick Check:

   Requests = 200Mi, Limits = 400Mi [OK]

Hint: Requests = min guaranteed, limits = max allowed [OK]

Common Mistakes:

• Swapping requests and limits values
• Setting only requests or only limits
• Setting requests higher than limits