DEPRECATED

Kubernetes Monitor service is no longer public available.

Kubernetes Monitoring Collector

This is a monitoring "collector" pod running in your cluster and reading info from the Kubernetes API and metrics API.

The data is sent to a Kubernetes Monitor Service. There you can configure a web dashboard and configure alarms (missing data, pod or container error status) and send notifications (email and/or webhook per namespace) to your DevOps teams.

Detailed docu: https://www.kubernetes-monitor.com/

TL;DR

You need to generate credentials for a "Monitoring Dashboard" per cluster.

"Free Tier" credentials unlock all feature and has no time limitation:

https://lcs.online-service.cloud/index.html?layout=product-nonav&id=613c6222be1a810011a01665

Install the "collector" pod:

kubectl create namespace monitoring 
kubectl apply -f mon-collector-rbac.yml -n monitoring 
kubectl create secret generic monitoring-secret -n monitoring \
      --from-literal=id="$MONITORING_ID" \
      --from-literal=key="$MONITORING_KEY" 
kubectl apply -f mon-collector.yml -n monitoring 

Open the dashboard https://www.kubernetes-monitor.com/ and configure services to monitor.

Install Collector As Pod (from Docker Hub)

The original config files and here: https://github.com/ma-ha/k8s-mon-collector

To get the files we recommend to clone this repo:

git clone https://github.com/ma-ha/k8s-mon-collector.git
cd k8s-mon-collector

Step 1: Create a new namespace

kubectl create namespace monitoring 

Step 2: Setup RBAC

The collector pod should only have read access to the Kubernetes API server:

kubectl apply -f mon-collector-rbac.yml -n monitoring 

Step 3: Configure Credentials

Copy/paste the command from the "Setup" tab in the Service Portal, should look like this:

kubectl create secret generic monitoring-secret -n monitoring \
  --from-literal=id="XXXXXXXXXXXXXX" \
  --from-literal=key="YYYYYYYYYYYYYYYY" 

Step 4: Deploy the Collector Pod

Only one pod is required:

kubectl apply -f mon-collector.yml -n monitoring 

Check the logs, there should be no errors.

kubectl get pods-n monitoring

Status should be "Running"

kubectl logs mon-collector-XXXXXXXX-XXXXX -n monitoring

Should look like:

info Starting mon-collector vX.Y.Z NODE_ENV=PRD 
info KubeConfig loadFromCluster... 

The deployment should be compliant to best practice security policies.

The Collector Service is open source, you're welcome to review the source code.

REMARK: Kubernetes v1.20 (or older) may not work with k8s-mon-collector:1.5.0 (or higher) version. You can edit the mon-collector.yml and use k8s-mon-collector:1.4.8 instead.

Done :-)

Navigate to https://www.kubernetes-monitor.com/, login and configure the dashboard.

Don't forget to configure alarms (E-Mail or Webhook) in the Service Portal.

Update the Collector Version

Checkout CHANGELOG.md for fixes and improvements.

To update the collector in your cluster, simply do:

cd k8s-mon-collector
git pull
kubectl apply -f mon-collector.yml -n monitoring 

Easy ... again check the pod status and the logs (see above).

Build Private Registry Image

To build a collector docker image an upload it to your private registry, please follow the instructions in README-BUILD.md.

Don't forget to change the container image in mon-collector.yml.

License

This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.

https://creativecommons.org/licenses/by-nd/4.0/