Chaos experiments on sock-shop app with grafana dashboard to monitor it.
-
Clone the sock-shop repo
git clone https://github.com/ksatchit/sock-shop.git cd sock-shop
-
Create sock-shop namespace on the cluster
kubectl create ns sock-shop
-
Apply the sock-shop microservices manifests
kubectl apply -f deploy/sock-shop/
-
Wait until all services are up. Verify via
kubectl get pods -n sock-shop
-
Install the litmus chaos operator and CRDs
kubectl apply -f https://litmuschaos.github.io/litmus/litmus-operator-v1.5.0.yaml
-
Install the litmus-admin serviceaccount for centralized/admin-mode of chaos execution
kubectl apply -f https://litmuschaos.github.io/litmus/litmus-admin-rbac.yaml
-
Install the chaos experiments in admin(litmus) namespace
kubectl apply -f https://hub.litmuschaos.io/api/chaos/1.5.0?file=charts/generic/experiments.yaml -n litmus
-
Install the chaos experiment metrics exporter and chaos event exporter
kubectl apply -f deploy/litmus-metrics/01-event-router-cm.yaml kubectl apply -f deploy/litmus-metrics/02-event-router.yaml kubectl apply -f deploy/litmus-metrics/03-chaos-exporter.yaml
-
Apply the monitoring manifests in specified order
kubectl apply -f deploy/monitoring/01-monitoring-ns.yaml kubectl apply -f deploy/monitoring/02-prometheus-rbac.yaml kubectl apply -f deploy/monitoring/03-prometheus-configmap.yaml kubectl apply -f deploy/monitoring/04-prometheus-alert-rules.yaml kubectl apply -f deploy/monitoring/05-prometheus-deployment.yaml kubectl apply -f deploy/monitoring/06-prometheus-svc.yaml kubectl apply -f deploy/monitoring/07-grafana-deployment.yaml kubectl apply -f deploy/monitoring/08-grafana-svc.yaml
-
Access the grafana dashboard via the NodePort (or loadbalancer) service IP or via a port-forward operation on localhost
Note: To change the service type to Loadbalancer, perform a
kubectl edit svc prometheus -n monitoring
and replacetype: NodePort
totype: LoadBalancer
kubectl get svc -n monitoring
Default username/password credentials:
admin/admin
-
Add the prometheus datasource for Grafana via the Grafana Settings menu
-
Import the grafana dashboard "Sock-Shop Performance" provided here
-
For the sake of illustration, let us execute a CPU hog experiment on the
catalogue
microservice & a Memory Hog experiment on theorders
microservice in a staggered mannerkubectl apply -f chaos/catalogue/catalogue-cpu-hog.yaml
Wait for ~60s
kubectl apply -f chaos/orders/orders-memory-hog.yaml
-
Verify execution of chaos experiments
kubectl describe chaosengine catalogue-cpu-hog -n litmus kubectl describe chaosengine orders-memory-hog -n litmus