See the documentation on how to deploy Sock Shop using Minikube.
There are 2 sets of manifests for deploying Sock Shop on Kubernetes: one in the manifests directory, and complete-demo.yaml. The complete-demo.yaml is a single file manifest made by concatenating all the manifests from the manifests directory, so please regenerate it when changing files in the manifests directory.
All monitoring is performed by prometheus. All services expose a /metrics endpoint. All services have a Prometheus Histogram called request_duration_seconds, which is automatically appended to create the metrics _count, _sum and _bucket.
The manifests for the monitoring are spread across the manifests-monitoring and manifests-alerting directories.
To use them, please run kubectl create -f <path to directory>.
- Sock-shop grafana dashboards
- Alertmanager with 500 alert connected to slack
- Prometheus with config to scrape all k8s pods, connected to local alertmanager.
Grafana will be exposed on the NodePort 31300 and Prometheus is exposed on 31090. If running on a real cluster, the easiest way to connect to these ports is by port forwarding in a ssh command:
ssh -i $KEY -L 3000:$NODE_IN_CLUSTER:31300 -L 9090:$NODE_IN_CLUSTER:31090 ubuntu@$BASTION_IP
Where all the pertinent information should be entered. Grafana and Prometheus will be available on http://localhost:3000 or :9090.
If on Minikube, you can connect via the VM IP address and the NodePort.