Restart an Application Deployment
Restart an Application Deployment
During the deployment of a KeptnApp, it might be that the deployment fails due to an unsuccessful pre-deployment
evaluation or pre-deployment task.
This could happen because of, e.g., a misconfigured target value of a KeptnEvaluationDefinition, or a wrong URL being
checked in a pre deployment check.
To retry a KeptnApp deployment without incrementing the version of the KeptnApp, we introduced the concept of **
revisions** for a KeptnAppVersion.
This means that
whenever the spec of a KeptnApp changes, even though the version stays the same, the KLT Operator will create a new
revision of the KeptnAppVersion referring to the KeptnApp.
This way, when a KeptnApp failed due to a misconfigured pre-deployment check, you can first fix the configuration of
the KeptnTaskDefinition/KeptnEvaluationDefinition, then
increase the value of spec.revision of the KeptnApp and finally apply the updated KeptnApp manifest.
This will result in a restart of the KeptnApp.
Afterwards, all related KeptnWorkloadInstances will automatically refer to the newly
created revision of the KeptnAppVersion to determine whether they are allowed to enter their respective deployment
phase.
To illustrate this, let’s have a look at the following example:
apiVersion: v1
kind: Namespace
metadata:
name: restartable-apps
annotations:
keptn.sh/lifecycle-toolkit: "enabled"
---
apiVersion: lifecycle.keptn.sh/v1alpha2
kind: KeptnApp
metadata:
name: podtato-head
namespace: restartable-apps
spec:
version: "0.1.1"
revision: 1
workloads:
- name: podtato-head-entry
version: "0.1.2"
preDeploymentTasks:
- pre-deployment-check
---
apiVersion: lifecycle.keptn.sh/v1alpha2
kind: KeptnTaskDefinition
metadata:
name: pre-deployment-check
namespace: restartable-apps
spec:
function:
inline:
code: |
console.error("I failed")
process.exit(1)
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: podtato-head-entry
namespace: restartable-apps
labels:
app: podtato-head
spec:
selector:
matchLabels:
component: podtato-head-entry
template:
metadata:
labels:
component: podtato-head-entry
keptn.sh/workload: podtato-head-entry
keptn.sh/app: podtato-head
keptn.sh/version: "0.1.2"
spec:
terminationGracePeriodSeconds: 5
containers:
- name: server
image: ghcr.io/podtato-head/entry:0.2.7
imagePullPolicy: Always
ports:
- containerPort: 9000
env:
- name: PODTATO_PORT
value: "9000"
In this example, the KeptnApp executes a pre-deployment check which clearly fails due to the pre-deployment-check
task, and will therefore not be able to proceed with the deployment.
After applying this manifest, you can inspect the status of the created KeptnAppVersion:
$ kubectl get keptnappversions.lifecycle.keptn.sh -n restartable-apps
NAME APPNAME VERSION PHASE
podtato-head-0.1.1-1 podtato-head 0.1.1 AppPreDeployTasks
You will notice that the KeptnAppVersion will stay in the AppPreDeployTasks phase for a while, due to the pre-check
trying to run until a certain failure threshold is reached.
Eventually, you will find the KeptnAppVersion’s PredeploymentPhase to be in a Failed state, with the remaining
phases being Deprecated.
$ kubectl get keptnappversions.lifecycle.keptn.sh -n restartable-apps -owide
NAME APPNAME VERSION PHASE PREDEPLOYMENTSTATUS PREDEPLOYMENTEVALUATIONSTATUS WORKLOADOVERALLSTATUS POSTDEPLOYMENTSTATUS POSTDEPLOYMENTEVALUATIONSTATUS
podtato-head-0.1.1-1 podtato-head 0.1.1 AppPreDeployTasks Failed Deprecated Deprecated Deprecated Deprecated
Now, to fix the deployment of this application, we first need to fix the task that has failed earlier.
To do so, edit the pre-deployment-check KeptnTaskDefinition to the
following (kubectl -n restartable-apps edit keptntaskdefinitions.lifecycle.keptn.sh pre-deployment-check):
apiVersion: lifecycle.keptn.sh/v1alpha2
kind: KeptnTaskDefinition
metadata:
name: pre-deployment-check
namespace: restartable-apps
spec:
function:
inline:
code: |
console.error("Success")
After we have done that, we can restart the deployment of our KeptnApplication by incrementing the spec.revision
field by one
(kubectl -n restartable-apps edit keptnapps.lifecycle.keptn.sh podtato-head):
apiVersion: lifecycle.keptn.sh/v1alpha2
kind: KeptnApp
metadata:
name: podtato-head
namespace: restartable-apps
spec:
version: "0.1.1"
revision: 2 # Increased this value from 1 to 2
workloads:
- name: podtato-head-entry
version: "0.1.2"
preDeploymentTasks:
- pre-deployment-check
After those changes have been made, you will notice a new revision of the podtato-head KeptnAppVersion:
$ kubectl get keptnappversions.lifecycle.keptn.sh -n restartable-apps
NAME APPNAME VERSION PHASE
podtato-head-0.1.1-1 podtato-head 0.1.1 AppPreDeployTasks
podtato-head-0.1.1-2 podtato-head 0.1.1 AppDeploy
As you will see, the newly created revision podtato-head-0.1.1-2 has made it beyond the pre-deployment check phase and
has reached its AppDeployPhase.
You can also verify the execution of the pre-deployment-check by retrieving the list of KeptnTasks in
the restartable-apps namespace:
$ kubectl get keptntasks.lifecycle.keptn.sh -n restartable-apps
NAME APPNAME APPVERSION WORKLOADNAME WORKLOADVERSION JOB NAME STATUS
pre-pre-deployment-check-49827 podtato-head 0.1.1 klc-pre-pre-deployment-check--77601 Failed
pre-pre-deployment-check-65056 podtato-head 0.1.1 klc-pre-pre-deployment-check--57313 Succeeded
You will notice that for both the KeptnAppVersions and KeptnTasks the previous failed instances are still available,
as this might be useful historical data to keep track of
what went wrong during earlier deployment attempts.
