In the next environment, we have a inventory app that is running on k8s, the hypothesis is that our customers are able to get products with no problem even if our app gets scale down or up.
Create a new chaos experiment file called inventory-config.hcl
app = "inventory app"
description = "Chaos Experiment for inventory app, check if customers are able to get products in cart"
job "kubernetes" "pod" {
namespace = "inventory-app"
config {
tag = "app:inventory-prod"
chaos = "terminate"
count = 5
}
}
The previous job will find 5 pods with label app:inventory-prod and terminate those.
In the same experiment, let’s check how does the cluster behaves with such rapid changes
job "kubernetes" "deployment" {
namespace = "inventory-app"
config {
tag = "app:inventory-prod"
chaos = "update"
count = 20
}
}
This job will increase to 20 the number of pods of the deployment with label app:inventory-prod