v0.1.7

Charts affected

• k8s-service

Description

Add the ability to configure tmpfs mounts using the scratchPaths input value.

Related links

• #91

v0.1.6

Charts affected

• k8s-service

Description

Add the ability to configure securityContext at the pod level using podSecurityContext input value.

Special thanks

Special thanks to @RyuCaelum for their contribution for this feature!

Related links

• #88

v0.1.5

Charts affected

• k8s-service

Description

You can now configure custom resources that are not managed by the helm chart. Refer to the updated docs for more information.

Special thanks

Special thanks to @paul-pop for their initial contribution for this feature!

Related links

• #86
• #82

v0.1.4

Charts affected

• k8s-service

Description

You can now set persistent volumes on the Deployment using the new input persistentVolumes.

Special thanks

Special thanks to @austinphilp for their contribution!

Related links

• #80

v0.1.3

Charts affected

• k8s-service

Description

You can now set additional labels on the Deployment and Pod using the additionalDeploymentLabels and additionalPodLabels inputs.

Special thanks

Special thanks to @TRReeve for their contribution!

Related links

• #75

v0.1.2

Charts affected

• k8s-service

Description

Fix bug where imagePullSecrets was not correctly rendered on the ServiceAccount resource.

Special thanks

Special thanks to @paul-pop for their contribution!

Related links

• #72

v0.1.1

Charts affected

• k8s-service

Description

You can now configure environment variables from arbitrary sources using the additionalContainerEnv input value.

Special thanks

Special thanks to @jjneely and @Dhertz for their contribution!

Related links

• #60
• #67

v0.1.0: Canary support

Charts affected

• k8s-service [BACKWARDS INCOMPATIBLE]

Description

This release introduces support for deploying a canary deployment to test new versions of your app. Refer to the updated README for more information.

Migration guide

To use canaries, the existing Deployment needs to be updated to watch for different set of labels compared to the canary Deployment. Unfortunately, updating selector labels is an unsupported operation in Kubernetes and helm does not handle this transition gracefully.

To support this, the Deployment resource needs to be recreated so that they get the new labels. The easiest way to handle this is to delete the Deployment resource (using kubectl delete deployment) and let helm recreate it during the upgrade process. Note that this means you will have downtime as the Pods are being deleted and recreated, despite any pod disruption budgets.

If you wish to avoid downtime, you can perform a blue-green deployment by recreating the Deployment resource under a new name. You can do this with the following approach:

# Retrieve the configuration for the deployment created by helm.
kubectl get deployment $DEPLOYMENT_NAME -n $DEPLOYMENT_NAMESPACE -o yaml > temp.yaml

# Open temp.yaml and update the name of the Deployment so that it can be created alongside the old one.

# Apply the updated yaml file to create a temporary Deployment object under a different name and wait for rollout.
kubectl apply -f temp.yaml
kubectl rollout status deployments $DEPLOYMENT_NAME -n $DEPLOYMENT_NAMESPACE

# Delete the old deployment so that the chart will recreate it.
kubectl delete deployment $DEPLOYMENT_NAME -n $DEPLOYMENT_NAMESPACE

# Rollout the update. This should recreate the Deployment resource with the new selector labels
helm upgrade --wait $RELEASE_NAME gruntwork/k8s-service

# At this point, it is safe to delete the temporary deployment resource we created
kubectl delete -f temp.yaml

Special thanks

Special thanks to @zackproser for their contribution!

Related links

• #54
• #45
• #56

v0.0.13

Charts affected

• k8s-service

Description

This release fixes a bug where the ServiceAccount annotations were not properly aligned with the metadata block.

Special thanks

Special thanks to @bouge for their contribution!

Related links

• #53

v0.0.12

Charts affected

• k8s-service

Description

• You can now optionally request to create the ServiceAccount directly from the chart, using the new serviceAccount.create parameter.
• You can now optionally configure horizontal pod autoscalers using the horizontalPodAutoscaler parameter.

Special thanks

Special thanks to @AechGG for their contribution!

Related links

• #50
• #51