Skip to main content

KaaS monitoring (experimental)

This component is marked as experimental, and it is not part of the reference SCS installation available at https://monitoring.scs.community.

Enable KaaS layer monitoring​

TODO: Provide clear instructions on how to enable KaaS layer monitoring.

Please check out the mvp-0 tag and find the related comments on what needs to be done in the monitoring values to enable KaaS monitoring in your Observer cluster.

KaaS metric importer​

To test the Monitoring of the KaaS layer use case, deploy the Kaas-metric-importer into the Observer cluster.

The Kaas-metric-importer is a simple service through which the KaaS software registers and unregisters newly created or deleted KaaS clusters in the Observer monitoring. This functionality enables the Observer monitoring to differentiate between KaaS clusters deleted intentionally and those that have stopped writing metrics to the Observer monitoring for any reason.

kubectl apply -f kaas/kaas-metric-importer.yaml

The Kaas-metric-importer uses an image built from https://github.com/m3dbx/prometheus_remote_client_golang. It has mounted configmap and based on configmap keys it pushes custom metric kaas with label cluster and value 1 into the thanos receiver.

Example of configmap:

data:
  workload-cluster: ""

It pushes metric e.g. kaas{cluster="workload-cluster"} 1 to the Observer. It is important to keep the configmap up-to-date with your KaaS offering. This is automated e.g. in the KaaS mock service below.

KaaS mock service​

To evaluate the Monitoring of the KaaS layer use case and view actual metrics in your Observer monitoring cluster, you can launch the KaaS mock service.

Put your Observer monitoring cluster kubeconfig into the ./kaas/manifests/ directory and name it observer-kubeconfig.yaml (or adjust kaas service ./kaas/app/config.py accordingly).

If you're utilizing the KinD Observer deployment outlined in this tutorial, collect the kubeconfig using the following command:

kind get kubeconfig --name observer > ./kaas/manifests/observer-kubeconfig.yaml

All KaaS mock service dependencies can be installed via the corresponding ./kaas/requirements.txt file. Installing them into a Python virtualenv is recommended.

cd kaas
python3 -m venv .venv  # Optional
source .venv/bin/activate  # Optional
# Install kaas dependencies
pip install -r requirements.txt

# Launch the KaaS mock service
make kaas

At this point, you should have the ability to access the KaaS mock service Swagger UI:

http://127.0.0.1:8080/kaas

  • Create KaaS cluster through Swagger UI: Create Cluster or call directly the KaaS service API via some client, e.g.:

    curl -X POST -H "Content-Type: application/json" http://127.0.0.1:8080/api/clusters/ -d '{"name": "kaas"}'

    Navigate to the KaaS Monitoring dashboard in the Observer monitoring. After a few minutes (approximately 4), your KaaS cluster should become visible. Click on the cluster box to dive into KaaS cluster dashboards at a more detailed level. Repeat the process to explore further and gain deeper insights.

    Note: The disk utilization expression for the Docker environment has not been adjusted, so you will encounter non-realistic numbers in the nodes/disk sections. However, the other sections should accurately reflect the reality.

  • Retrieve a list of all KaaS clusters and check their status. Swagger UI: Get List of Clusters or call directly the KaaS service API via some client, e.g.:

    curl -s -X GET -H 'accept: application/json' http://127.0.0.1:8080/api/clusters/

  • Get Kaas Cluster kubeconfig by its name through Swagger UI: Get Cluster kubeconfig or call directly the KaaS service API via some client and save it, e.g.:

    curl -s -X GET -H 'accept: application/json' http://127.0.0.1:8080/api/clusters/kaas > kaas-kube

  • Now, you have the opportunity to play with your KaaS cluster and experiment with triggering monitoring alerts by initiating actions like destroying certain components 😎.

    kubectl --kubeconfig kaas-kube get po -A

  • Finally, delete your KaaS cluster by its name through Swagger UI: Delete Cluster or call directly the KaaS service API via some client and save it, e.g.:

    curl -X DELETE http://127.0.0.1:8080/api/clusters/?name=kaas