052

README

Episode 52 : Instrumenting with Prometheus

• Hosted by @krisnova
• Recording date: 2018-10-05

Table of Contents

• 00:00:00 - Welcome to TGIK!
• 00:00:00 - Week in Review

K8s news

• Kubernetes 1.12 is out. 1.12.1 came out just now!
• Steering Committee Election Results - congrats to Aaron Crickenberger, Tim St. Clair, and Davanum Srinivas
• The folks from automaticserver have created a CRI shim for LXC/LXD
• Gardener .12 released
• Schedule for Kubecon is now available
• Teleport 3.0 is out
• Digital Ocean's k8s now in limited availability

Show Notes

Set up cluster with Prometheus

Create a cluster in AWS with Kubicorn

kubicorn create prom-cluster -paws
kubicorn apply prom-cluster

Install Prometheus and other apps

kubectl apply -f bundle.yaml
kubectl apply -f prometheus.yaml

Poke a hole in the firewall

• Open the Node security group
• Expose the TCP Ports to each LoadBalancer created
  • 9093 -> alertmanager
  • 3000 -> grafana
  • 9090 -> prometheus

Change the Makefile to fit your docker registry if needed, and build the example app

make container push

Deploy the example app

kubectl run tgik-app -n monitoring --image krisnova/tgik-instrumentation-app
kubectl expose -n monitoring tgik-app --port 9093 --target-port 1313

To reload the application

kubectl scale tgik-app -n monitoring --replicas=0
kubectl scale tgik-app -n monitoring --replicas=1

To reload Prometheus without taking it down:

curl -i -X POST localhost:9090/-/reload

Do not expose these lifecycle endpoints to the public

Reference Links

• Instrumenting Go with Prometheus
• Joe's Prometheus v1
• All the Autoscaling
• Step by Step CRDs
• Cloud Native London Videos
• Prometheus Query Language
• Original Source of Prometheus

https://github.com/prometheus/client_golang/blob/master/prometheus/go_collector.go#L272