Kubernetes AutoStopping for Istio
This article provides instructions to create AutoStopping rules for Kubernetes with the Istio ingress controller.
Prerequisites
Make sure to meet the following prerequisites before you create an AutoStopping rule for Istio.
- Make sure you are running at least version 1.0.8 of
autostopping-controller
in your Kubernetes cluster. - Istio gateway and virtual service are configured and are routing traffic to the service as expected.
- The latest cost optimization connector YAML applied in the cluster with the read-write permissions to
networking.istio.io object virtualservices
Set up your cluster
Create an AutoStopping rule in your cluster, either through the Harness UI or by applying YAML directly on the cluster of the form:
apiVersion: ccm.harness.io/v1
kind: AutoStoppingRule
metadata:
name: test-rule
namespace: default
annotations:
harness.io/cloud-connector-id: CloudAccountID
spec:
service:
name: echo
port: 8080
istio:
virtualServices:
- httpbin
idleTimeMins: 5
hideProgressPage: false
After applying the YAML, an AutoStopping Rule is created in your cluster for service echo which is running on port 8080.
Now, your traffic flows through autostopping-router, and the AutoStopping takes actions based on activity and inactivity.
Multiple VirtualServices Pointing to the same Service
If you have a setup where multiple virtual services are pointing to the same Kubernetes service, you can configure AutoStopping for those services as given in the following example.
Assume httpbin1
and httpbin2
are your Istio virtualservices and echo is your Kubernetes service.
apiVersion: ccm.harness.io/v1
kind: AutoStoppingRule
metadata:
name: test-rule
namespace: default
annotations:
harness.io/cloud-connector-id: CloudAccountID
spec:
service:
name: echo
port: 8080
istio:
virtualServices:
- httpbin1
- httpbin2
idleTimeMins: 5
hideProgressPage: false
After applying this rule, both of the virtualservices are reconfigured for AutoStopping orchestration.
VirtualServices Pointing to the Service in different Namespace
If you have a setup where a virtualservice is pointing to a Kubernetes service in a different namespace with its FQDN, you can configure AutoStopping for those services as given in the following example.
Assume httpbin1
is your Istio virtualservice in default namesapce and echo is your Kubernetes service in namespace development.
apiVersion: ccm.harness.io/v1
kind: AutoStoppingRule
metadata:
name: test-rule
namespace: development
annotations:
harness.io/cloud-connector-id: CloudAccountID
spec:
service:
name: echo
port: 8080
istio:
virtualServices:
- default.httpbin1
idleTimeMins: 5
hideProgressPage: false
Here, you are creating the rule in the same namespace as that of the Kubernetes service (development) and referring to the virtualservice in the default namespace. After applying this rule, the virtualservice is reconfigured for AutoStopping orchestration.