Metric Types

Question 1

list metric types

Answer 1

• counter,
  -gauge,
  -histogram,
• summary

Question 2

explain counter metric

Answer 2

feature that can only go up

Question 3

explain gauge metric

Answer 3

like counter but can also go down, not only up, for example room temperature

Question 4

explain downsides of calculating metrics on client side

Answer 4

• pressure moved from prometheus to the app,
• can’t chose the time window, there’s only data for last 10 minutes
• can’t aggregate quantiles because each pod/server will deliver it’s own metrics and averaging quantiles is mathematically incorrect

Question 5

Why averaging quantiles is mathematically incorrect?

Answer 5

imagine it takes me 1 day to take out trash and 9 other people it takes on 10 minutes. If I average this it’ll make no sense… Imagine people are instances.