Monitoring Kubernetes clusters on AWS, GCP and Azure using Prometheus Operator by CoreOS

Note: the work on this repository is now based on CoreOS's kube-prometheus and it will be the default option for Kubernetes 1.7.X and up. For 1.5.X and 1.6.X you can deploy a simpler solution, located in ./basic directory. The purpose of this project is to provide a simple and interactive method to deploy and configure Prometheus on Kubernetes, especially for the users that are not using Helm.

Features

• Prometheus Operator with support for Prometheus v2.X.X
• highly available Prometheus and Alertmaneger
• InCluster deployment using StatefulSets for persistent storage
• auto-discovery for services and pods
• automatic RBAC configuration
• preconfigured alerts
• preconfigured Grafana dashboards
• easy to setup; usually less than a minute to deploy a complete monitoring solution for Kubernetes
• support for Kubernetes v1.7.x and up running in AWS, GCP and Azure
• tested on clusters deployed using kube-aws, kops, GKE and Azure

One minute deployment

Prerequisites

• Kubernetes cluster and kubectl configured
• Security Groups configured to allow the following ports:
  • 9100/TCP - node-exporter
  • 10250/TCP - kubernetes nodes metrics,
  • 10251/TCP - kube-scheduler
  • 10252/TCP - kube-controller-manager
  • 10054/TCP and 10055/TCP - kube-dns

Optional

• SMTP Account for email alerts
• Token for Slack alerts

Running Kubernetes 1.12 and up?

If you are running Kubernetes 1.12 or higher you will also need to run cAdvisor on your cluster (bound to host port 4194) in order to access resource usage and performance characteristics of running containers.

Pre-Deployment

Clone the repository and checkout the latest release: curl -L https://git.io/getPrometheusKubernetes | sh -

Custom settings

All the components versions can be configured using the interactive deployment script. Same for the SMTP account or the Slack token.

Some other settings that can be changed before deployment:

• Prometheus replicas: default 2 ==> manifests/prometheus/prometheus-k8s.yaml
• persistent volume size: default 40Gi ==> manifests/prometheus/prometheus-k8s.yaml
• allocated memory for Prometheus pods: default 2Gi ==> manifests/prometheus/prometheus-k8s.yaml
• Alertmanager replicas: default 3 ==> manifests/alertmanager/alertmanager.yaml
• Alertmanager configuration: ==> assets/alertmanager/alertmanager.yaml
• custom Grafana dashboards: add yours in assets/grafana/ with names ending in -dashboard.json
• custom alert rules: ==> assets/prometheus/rules/

Note: please commit your changes before deployment if you wish to keep them. The deploy script will remove the changes on most of the files.

Deploy

./deploy

Now you can access the dashboards locally using kubectl port-forwardcommand, or expose the services using a ingress or a LoadBalancer. Please check the ./tools directory to quickly configure a ingress or proxy the services to localhost.

To remove everything, just execute the ./teardown script.

Updating configurations

• update alert rules: add or change the rules in assets/prometheus/rules/ and execute scripts/generate-rules-configmap.sh. Then apply the changes using kubectl apply -f manifests/prometheus/prometheus-k8s-rules.yaml -n monitoring
• update grafana dashboards: add or change the existing dashboards in assets/grafana/ and execute scripts/generate-dashboards-configmap.sh. Then apply the changes using kubectl apply -f manifests/grafana/grafana-dashboards.cm.yaml.

Note: all the Grafana dashboards should have names ending in -dashboard.json.

Custom Prometheus configuration

The official documentation for Prometheus Operator custom configuration can be found here: custom-configuration.md If you wish, you can update the Prometheus configuration using the ./tools/custom-configuration/update_config script.