Upgrade
If you installed the previous version of Keptn using helm
,
you can upgrade to the latest version
by running the same command sequence used to install Keptn:
helm repo add keptn https://charts.lifecycle.keptn.sh
helm repo update
helm upgrade --install keptn keptn/keptn \
-n keptn-system --create-namespace --wait
Use the --set
flag or download and edit the values.yaml
file
to modify the configuration as discussed on the
Install Keptn page.
Warning If you installed your Keptn instance from the Manifest, additional steps are required to use the Helm Chart to upgrade. Contact us on Slack for assistance.
Upgrade to Helm from a manifest installation
Warning Upgrade to Helm from a manifest installation can cause loss of Keptn custom resources.
Keptn v.0.7.0 and later can be installed with Helm charts; Keptn v.0.8.3 can only be installed with Helm. If you previously installed Keptn from manifests, you can not directly upgrade with Helm but must back up your manifests, then reinstall Keptn from a Helm chart and re-apply your manifests.
To start the upgrade process, follow the steps below:
-
To not loose all of your data, we encourage you to do a backup of the manifests, which you applied to the cluster (
Pods
,Deployments
,StatefulSets
,DaemonSets
,KeptnApps
,… ). After the re-installation of Keptn with Helm, you can re-apply these manifests and restart the Keptn deployment process. -
Completely remove your Keptn installation with the following command sequence:
your-keptn-version=<your-keptn-version>
kubectl delete -f \
https://github.com/keptn/lifecycle-toolkit/releases/download/$your-keptn-version/manifest.yaml
- Use Helm to install a clean version of Keptn:
helm repo add keptn https://charts.lifecycle.keptn.sh
helm repo update
helm upgrade --install keptn keptn/keptn -n keptn-system --create-namespace --wait
For information about advanced installation options, refer to Modify Helm configuration options.
- After the installation finishes, restore the manifests from you backup
Warning Please be aware that Keptn applications will start the deployment process from the beginning and the system is not guaranteed to return to the exact state it was in before re-installation, even if you created the backup correctly.
Migrate from v0.6.0 to v0.7.0
Keptn Version v0.7.0
introduces the metrics-operator
,
which is now separate from the lifecycle-operator
.
Some functionality and behavior has been moved, changed, or renamed.
Specifically, the KeptnMetricsProvider
CRD replaces
the now-deprecated KeptnEvaluationProvider
CRD.
Consequently, you must manually migrate the existing functionality
to the KeptnMetricsProvider
CRD.
Execute the following external bash script to do this:
curl -sL https://raw.githubusercontent.com/keptn/lifecycle-toolkit/main/.github/scripts/keptnevaluationprovider_migrator.sh | bash
This fetches and migrates all KeptnEvaluationProvider
CRs
for the cluster at which your kubernetes context is pointing.
If you have multiple clusters,
you must run this command for each one.
This script does the following:
- Fetch all existing
KeptnEvaluationProvider
CRs - Migrate each to
KeptnMetricsProvider
CRs - Store the migrated manifests in the current working directory
- Ask the user to apply the created manifests
and delete the deprecated
KeptnEvaluationProvider
CRs.