< Previous Challenge - Home - Next Challenge >
This challenge will cover monitoring in AKS, using open source components such as Prometheus and Azure services such as Azure Monitor.
- Implement Prometheus/Grafana
- You can access container logs via Azure Monitor
- Increase the CPU utilization of the API container with the
pi
API endpoint, and see the corresponding metric increase in Prometheus and/or Azure Monitor - Implement a mechanism so that Kubernetes increases the amount of API pods when CPU utilization goes high
- If you didn't do it already, configure a mechanism that scales the cluster automatically in and out depending on the required capacity
- You can display cluster metrics graphically
- You can show live container logs with Azure Container Insights
- Verify that the cluster auto-scales when there are not enough CPU resources
- Participants can explain the autoscaling event using AKS metrics
These docs might help you achieving these objectives:
Using the API of the solution you can create CPU utilization with these commands, that leverage the pi
endpoint of the
API (calculate pi number with x digits).
In the Challenge Resources directory you can find a bash script to generate load by repeatedly invoking the pi
endpoint.
../Challenge_Resources/Challenge_03/api-load-test.sh
You can check the increased CPU utilization in Azure Monitor in Prometheus.