exercise-3

Exercise 3

Introduction

Even if you know how to fix a memory leak, you would have preferred not having to!

Monitoring your application is the key.

In this exercise, you will start a Prometheus, make it scrape your application, and create an alert.

Start

Run the exercise application

npm run start-exercise-3

In another terminal, run the docker image "prom/prometheus"

docker run \
  --network host \
  --volume "/prometheus" \
  --volume "$(pwd)/src/exercise-3/prometheus-data:/prometheus-data" \
  prom/prometheus \
    --storage.tsdb.path=/prometheus-data/storage \
    --config.file=/prometheus-data/prometheus.yml 

Open http://localhost:9090, you should see the prometheus interface

Graphing

Prerequisites

Have a look at the prometheus queries!

Watch the heap!

From there, create graphs showing the heap spaces sizes over 5 minutes.

Solution

• click on "add graph" to get 2 graphs
• switch to graph mode for both
• enter the following queries

  nodejs_heap_space_size_used_bytes{space="new"}
  

  and

  nodejs_heap_space_size_used_bytes{space="old"}
  

• press "execute" for both

Watch the GC!

Create another graph on the same page showing the garbage collector reclaimed bytes increase per minute over 5 minutes..

Solution

• click on "add graph"
• switch to graph mode
• enter the following query

  increase(nodejs_gc_reclaimed_bytes_total{}[1m])
  

• check "stacked"
• press "execute"

Eventually, you should reach this state.