Observability - Monitor and Debug Applications

Question 1

What do we usually want to monitor?

Answer 1

• node level metrics: how many nodes
• performance metrics: cpus, memory
• pod level metrics: nr of pods, consumption

Question 2

What can we use to monitor in Kubernetes?

Answer 2

No built-in solution yet.
Many open source solutions: MetricsServer, Prometheus, Elastic Stack

Question 3

What is Metrics Server? How can you work with it?

Answer 3

• logging solution
• one metric server per kubernetes cluster

Question 4

What does a Metrics Server do?

Answer 4

• retrieves metrics from each of the Kubernetes nodes and pods
• aggregates and stores them in memory
• only in-memory storing solution
• no historical data
• for that advanced monitoring solution

Question 5

How does the kubelet work for logging?

Answer 5

• has a sub component cAdvisor
• cAdvisor is responsible for retrieving performance metrics from pods
• and exposing them through kubelet API to make them available for the metrics server

Question 6

How can you deploy a metrics server?

Answer 6

1. git clone https://github.com/kubernetes-incubator/metrics-server.git
2. ‘kubectl create -f deploy/1.8+/’
3. kubectl top node -> to see the collected data