Kubernetes Resource Lecture

Question 1

In Kubernetes, what resources does each node provide for pods?

Answer 1

CPU and memory.

Question 2

What does each pod require in order to run on a node?

Answer 2

A specified amount of CPU and memory resources.

Question 3

Which Kubernetes component decides which node a pod is placed on?

Answer 3

The Kubernetes Scheduler.

Question 4

What does the Kubernetes scheduler consider when placing a pod on a node?

Answer 4

The pod’s resource requests and the available resources on each node.

Question 5

What happens if no node has sufficient resources to satisfy a pod’s request?

Answer 5

The pod remains in the Pending state and is not scheduled.

Question 6

In a Pending pod, what might you see in kubectl describe pod if there aren’t enough CPU resources?

Answer 6

An event indicating insufficient CPU (or similar resource-related message).

Question 7

In Kubernetes, what is a resource request for a container?

Answer 7

The minimum amount of CPU or memory that a container asks for and is guaranteed.

Question 8

Where do you define resource requests in a pod manifest?

Answer 8

Under spec.containers[].resources.requests for each container.

Question 9

How does the scheduler use resource requests?

Answer 9

It uses them to find a node with enough free capacity to run the pod.

Question 10

What is the effect of setting a CPU request for a container?

Answer 10

The container is guaranteed at least that amount of CPU when there is contention.

Question 11

What is the effect of setting a memory request for a container?

Answer 11

The container is guaranteed at least that amount of memory when there is contention.

Question 12

In Kubernetes, what does 1 CPU represent?

Answer 12

Roughly 1 vCPU/core/hyperthread (e.g., 1 vCPU in AWS or 1 core in GCP/Azure).

Question 13

How can you represent 0.1 CPU using milliCPU units?

Answer 13

As 100m (100 millicpu).

Question 14

What is the smallest CPU value you can request in Kubernetes?

Answer 14

1m (1 millicpu).

Question 15

Can you request non-integer CPU amounts like 0.3 CPU?

Answer 15

Yes, for example 0.3 CPU ≈ 300m.

Question 16

What is the difference between G and Gi when specifying memory?

Answer 16

G is gigabyte (GB, base 1000); Gi is gibibyte (GiB, base 1024).

Question 17

What does 256Mi represent?

Answer 17

256 mebibytes (binary unit, base 1024).

Question 18

Which suffix family is base-10 for memory units in Kubernetes?

Answer 18

K, M, G (e.g., MB, GB).

Question 19

Which suffix family is base-2 for memory units in Kubernetes?

Answer 19

Ki, Mi, Gi (e.g., MiB, GiB).

Question 20

In Kubernetes, what is a resource limit for a container?

Answer 20

The maximum amount of CPU or memory that a container is allowed to use.

Question 21

Where do you set resource limits in a pod manifest?

Answer 21

Under spec.containers[].resources.limits.

Question 22

Are resource requests and limits applied per pod or per container?

Answer 22

Per container; each container has its own requests and limits.

Question 23

What happens if a container tries to use more CPU than its limit?

Answer 23

The CPU usage is throttled so it cannot exceed its CPU limit.

Question 24

Does exceeding CPU limits by itself cause a container to be killed?

Answer 24

No, CPU overuse is throttled rather than killing the container.

Question 25

What happens if a container uses more memory than its limit?

Answer 25

It can be killed by the OOM (Out Of Memory) killer (OOMKilled).

Question 26

Why can CPU usage be throttled but memory usage cannot?

Answer 26

CPU can be scheduled and limited over time; memory, once allocated, can only be reclaimed by freeing it, often by killing the pod.

Question 27

By default, does Kubernetes set resource requests for pods?

Answer 27

No, by default there are no resource requests set.

Question 28

By default, does Kubernetes set resource limits for pods?

Answer 28

No, by default there are no resource limits set.

Question 29

What is the risk of running pods with no resource requests and no limits?

Answer 29

One pod can consume all CPU or memory on a node, starving other pods and system processes.

Question 30

If two pods have no CPU requests and no limits, how is CPU shared?

Answer 30

Unfairly; one pod can hog all CPU while the other gets very little.

Question 31

If CPU limits are set but requests are not, what does Kubernetes do?

Answer 31

It automatically sets CPU requests equal to the CPU limits.

Question 32

In the CPU scenario where only limits are set, what does each pod effectively get?

Answer 32

Each pod is guaranteed CPU equal to its limit and cannot exceed that limit.

Question 33

In the CPU scenario where requests=1 and limits=3, what does that mean for a pod?

Answer 33

The pod is guaranteed 1 vCPU and can use up to 3 vCPU but no more.

Question 34

Why might “requests and limits both set” for CPU be suboptimal?

Answer 34

A pod that could use more CPU is capped at its limit even if the node has spare CPU and other pods are idle.

Question 35

Why is “CPU requests set, but no CPU limits” often considered ideal?

Answer 35

Each pod has a guaranteed minimum CPU, and pods can use extra CPU if it’s available instead of being capped.

Question 36

If two pods have no memory requests and no limits, what can happen?

Answer 36

One pod can consume all memory, causing the other to be starved or killed.

Question 37

If memory limits are set but requests are not, what does Kubernetes do?

Answer 37

It automatically sets memory requests equal to the memory limits.

Question 38

In the memory scenario with requests=1Gi and limits=3Gi, what happens if a pod exceeds 3Gi?

Answer 38

It can be OOMKilled for exceeding its memory limit.

Question 39

What is the main risk of setting memory requests but no memory limits?

Answer 39

A pod can grow to use all available memory, causing OOMKills and node instability.

Question 40

What is a LimitRange in Kubernetes?

Answer 40

A namespace-scoped object that defines default requests/limits and min/max constraints for containers in pods.

Question 41

At what scope is a LimitRange applied?

Answer 41

At the namespace level.

Question 42

What can a LimitRange configure for CPU or memory?

Answer 42

Default limits, default requests, max limits, and min requests.

Question 43

When does a LimitRange take effect on pods?

Answer 43

Only for new pods created or updated after the LimitRange is applied or changed.

Question 44

Does changing a LimitRange affect existing pods immediately?

Answer 44

No, existing pods are not changed; only new pods are affected.

Question 45

What is a ResourceQuota in Kubernetes?

Answer 45

A namespace-scoped object that sets hard limits on the total resource usage in that namespace.

Question 46

At what scope is a ResourceQuota applied?

Answer 46

At the namespace level.

Question 47

What can a ResourceQuota limit with respect to CPU and memory?

Answer 47

Total requested CPU/memory and total CPU/memory limits for all pods in a namespace.

Question 48

Why would you use a ResourceQuota?

Answer 48

To prevent one namespace or team from consuming too many cluster resources and to enforce fair multi-tenant usage.

Question 49

What is a common recommended pattern for CPU configuration in Kubernetes?

Answer 49

Set CPU requests for all containers and optionally avoid CPU limits to allow bursting when capacity is available.

Question 50

What is a key concern when skipping memory limits for containers?

Answer 50

A misbehaving pod can consume all memory and cause OOMKills and node instability.

Question 51

Why is it important that all pods have some CPU or memory requests set, especially if you skip limits?

Answer 51

Requests guarantee minimum resources; without them, any pod with no request can be starved by pods that consume all available CPU or memory.