Observability

Question 1

What is the imperative command to view the current Pod status for all pods?

Answer 1

kubectl get po

Question 2

What are Pod conditions?

Answer 2

Additional information related to the Pod status. Array of boolean values, as to whether Pod is Scheduled, Initialized, ContainersReady, and Pod is Ready.

Question 3

What is the imperative command to view the Pod conditions for a Pod? (e.g. for Pod po1)

Answer 3

kubectl describe po po1

View the Conditions section

Question 4

What is the purpose of the Readiness probe?

Answer 4

Prevents service disruption while Pods are scaled up. The Readiness probe provides kubernetes with additional information about the state of the application running in a Pod. Kubernetes can use this to determine whether to send traffic to a Pod.

Question 5

What are three ways of determining readiness of a Pod?

Answer 5

• HTTP request to the Pod
• TCP request to the Pod
• Execute a Script within a container of the Pod

Question 6

In the Pod definition, how is an HTTP readiness probe declared? (e.g. HTTP on Port 80 for path /example/path)

Answer 6

spec: containers - readinessProbe: httpGet: path: /example/path port: 80

Question 7

In the Pod definition, how is an TCP readiness probe declared? (e.g. TCP on Port 3000)

Answer 7

spec: containers: - readinessProbe: tcpSocket: port: 3000

Question 8

In the Pod definition, how is an executable readiness probe declared? (e.g. cat file /var/log/ready)

Answer 8

spec: containers: - readinessProbe: exec: command: - cat - /var/log/ready

Question 9

In the Pod definition, how are initial delay and how often to repeat a readiness probe defined? (e.g. 10 seconds for initial delay and 5 seconds for repetition)

Answer 9

spec: containers: - readinessProbe: initialDelaySeconds: 10 periodSeconds: 5

Question 10

In the Pod definition, how is the maximum number of failures of the readiness probe defined? (e.g. 3 failures)

Answer 10

spec: containers: - readinessProbe: failureThreshold: 3

Question 11

What is the purpose of the liveness probe?

Answer 11

Prevents traffic being sent to a probe that is not healthy. The liveness probe provides kubernetes with additional information about the state of the application running in a Pod. Kubernetes can use this to determine whether to send traffic to a Pod.

Question 12

In the Pod definition, how is an HTTP liveness probe declared? (e.g. HTTP on Port 80 for path /api/healthy)

Answer 12

spec containers: - livenessProbe: httpGet: path: /api/healthy port: 80

Question 13

In the Pod definition, how is an TCP readiness probe declared? (e.g. TCP on Port 3001)

Answer 13

spec: containers: - livenessProbe: tcpSocket: port: 3001

Question 14

In the Pod definition, how is an executable readiness probe declared? (e.g. cat file /var/log/healthy)

Answer 14

spec: containers: - livenessProbe: exec: command: - cat - /var/log/healthy

Question 15

In the Pod definition, how are initial delay and how often to repeat a liveness probe defined? (e.g. 10 seconds for initial delay and 5 seconds for repetition)

Answer 15

spec: containers: - livenessProbe: initialDelaySeconds: 10 periodSeconds: 5

Question 16

In the Pod definition, how is the maximum number of failures of the liveness probe defined? (e.g. 3 failures)

Answer 16

spec: containers: - livenessProbe: failureThreshold: 3

Question 17

What is the imperative command for viewing logs of a Pod with a single container? (e.g. Pod po1)

Answer 17

kubectl logs po1

Question 18

What is the imperative command for viewing logs of a Pod with multiple containers? (e.g. Container c1 of Pod po1)

Answer 18

kubectl logs po1 c1

Question 19

What is the purpose of Metrics Server?

Answer 19

It provides metrics data about Kubernetes Pods and Nodes. One metrics server can be deployed per cluster. It aggregates the metrics and stores in memory.

Question 20

What is one limitation of Metrics Server?

Answer 20

It stores data in memory, and does not provide persistence.

Question 21

How does the Metrics Server work?

Answer 21

The cAdvisor component of Kubelet, which is installed on each node, collects metrics. Kubelet exposes these metrics through its API which Metrics Server can send requests to.

Question 22

How is metrics server deployed to kubernetes?

Answer 22

kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

Question 23

What is the imperative command for viewing metrics about nodes?

Answer 23

kubectl top node

Question 24

What is the imperative command for viewing metrics about pods?

Answer 24

kubectl top po