Doc description
GitHub Repo: https://github.com/DevOps-TechLab/kube-prometheus-hpa-monitor
This Repo has the following two functions:
- monitor kubernetes
- kubernetes custom hpa monitor
How to monitor Aks/K8S
TIPS: The metrics-server is deployed by default in the aks cluster, and there is no need to deploy metrics-server again in the azure environment.
kubectl apply -f ./namespaces.yaml
kubectl apply -f ./metrics-server/0.3.6/
kubectl apply -f ./kube-state-metrics/
kubectl aply -f ./prometheus/
K8S HPA
Customize monitoring indicators through Prometheus adaptor:
https://github.com/stefanprodan/k8s-prom-hpa
How to depoy it:
We still need to perform the steps of deploying prometheus above, and perform the following steps after completion:
make certs
kubectl apply -f ./custom-metrics-api/
Prometheus federation
How does external prometheus grab monitor data of k8s-prometheus?
The prometheus pod is already running in the k8s cluster, and external access is provided through nodeport, we only need to configure remote crawling:
- job_name: 'kubernetes-cluster'
scrape_interval: 30s
honor_labels: true
metrics_path: '/federate'
params:
'match[]':
- '{job=~"kubernetes-.*"}'
static_configs:
- targets:
- k8s1-node1-ip:30090
- k8s1-node2-ip:30090
labels:
k8scluster: noprod-cluster
environment: 'non-production'
resourcetype: 'azure-paas'
department: 'BU1'
- targets:
- k8s2-node1-ip:30090
- k8s2-node2-ip:30090
labels:
k8scluster: prod-cluster
environment: 'production'
resourcetype: 'azure-paas'
department: 'BU1'
...