Set up Operational Rules
You can set up operational rules in the AppConfiguration
model via the opsRule
field and corresponding platform configurations in the workspace directory. The opsRule
is the collection of operational rule requirements for the application that are used as a preemptive measure to police and stop any unwanted changes.
Prerequisitesโ
Please refer to the prerequisites in the guide for deploying an application.
The example below also requires you to have initialized the project using the kusion workspace create
and kusion init
command, which will create a workspace and also generate a kcl.mod
file under the stack directory.
Managing Workspace Configurationโ
In the first guide in this series, we introduced a step to initialize a workspace with an empty configuration. The same empty configuration will still work in this guide, no changes are required there.
However, if you (or the platform team) would like to set default values for the opsRule to standardize the behavior of applications, you can do so by updating the ~/dev.yaml
:
modules:
opsRule:
default:
maxUnavailable: "40%"
Please note that the maxUnavailable
in the workspace configuration only works as a default value and will be overridden by the value set in the application configuration.
The workspace configuration need to be updated with the command:
kusion workspace update dev -f ~/dev.yaml
If the platform engineers have set the default workload to Kusion Operation and installed the Kusion Operation controllers properly, the opsRules
module will generate a PodTransitionRule instead of updating the maxUnavailable
value in the deployment
Exampleโ
Add the opsRule
snippet to the AppConfiguration
in simple-service/dev/main.k
:
import catalog.models.schema.v1 as ac
import catalog.models.schema.v1.workload as wl
import catalog.models.schema.v1.workload.container as c
import catalog.models.schema.v1.trait as t
helloworld: ac.AppConfiguration {
workload: wl.Service {
...
}
# Configure the maxUnavailable rule
opsRule = t.OpsRule {
maxUnavailable: "30%"
}
}
Applyingโ
Re-run steps in Applying, resource scaling is completed.
$ kusion apply
โ๏ธ Generating Intent in the Stack dev...
Stack: dev ID Action
* โโ v1:Namespace:simple-service UnChanged
* โโ v1:Service:simple-service:simple-service-dev-helloworld-private UnChanged
* โโ apps/v1:Deployment:simple-service:simple-service-dev-helloworld Update
? Do you want to apply these diffs? yes
Start applying diffs ...
SUCCESS UnChanged v1:Namespace:simple-service, skip
SUCCESS UnChanged v1:Service:simple-service:simple-service-dev-helloworld-private, skip
SUCCESS Update apps/v1:Deployment:simple-service:simple-service-dev-helloworld success
Update apps/v1:Deployment:simple-service:simple-service-dev-helloworld success [3/3] โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ 100% | 0s
Apply complete! Resources: 0 created, 1 updated, 0 deleted.
Validationโ
We can verify the application deployment strategy now has the updated attributes maxUnavailable: 30%
in the container configuration:
kubectl get deployment -n simple-service -o yaml
...
apiVersion: apps/v1
kind: Deployment
...
spec:
strategy:
rollingUpdate:
maxUnavailable: 30%
type: RollingUpdate
...