terraform-provider-kubectl

Fix handling of yaml fields to allow for multiple fields with the same key (e.g one at labels.name and spec.name)

Note: Due to this change, when upgrading you may notice a large amount of "changed" resources for kubectl_manifest stanzas. These are expected.

Changelog

6715529 Don't print deployment details on error as could leak sensitive content

Changelog

e8aca4e Allow disabling of schema validation
245e4de Doc tweaks

New! You can now define sensitive_fields on kubectl_manifest resources and the kubectl provider will obfuscate those fields for you like magic! 🧙

Changelog

666ea2b Sensitive field support to obfuscate values
89e888e Added goreleaser & skip travis releasing
b1d321c Added registry.terraform.io docs format

Fix panic on differing types for updated yaml manifests

🚀 Performance Release!

1. Uses a cached discovery client to allow reuse of the disk based kube cache
2. Added smarter handling of retry backoff to correctly handle managing applying manifests which create dependent resources
3. Added better logging when running with debug TF modes
4. Added timeout detector for creating a discovery client to ensure large projects don't time out

Remove waiting for DaemonSet's to rollout, causes issues with large clusters. Users should add their own waiting scripts as required.

Emit more details for failed deployments

Increase timeout back to 10mins for deployments/daemonset monitoring

• Fix issues with retry logic retrying too fast & buffer initial retry by 2 seconds
• Fix issues when retry being disabled is not respected
• Default timeout durations to 5mins for resources
• Upgrade Kuberentes API versions to v1.17.x
• Upgrade golang to v1.14

Handle delete case when kubernetes resource has already been deleted

Handle deletion of kubernetes resources by removing from state during refresh. This fixes an issue where resources may have been deleted by other processes and terraform would fail state refresh.

Fix panic when comparing some objects with invalid keys

Added triggers to kubectl_server_version to allow re-reading of the server version when upgrading

Added resource for kubectl_server_version to match existing data source introduced in v1.3.0

Added kubectl_server_version data source to fetch the servers version docs.

Fix #14 - panic when applying manifests on top of existing resources which were not created with kubectl apply, rather using kubectl create or other means

The data source kubectl_path_documents now supports terraform template evaluation performed during loading documents from file 🎉 This gives you the flexibility of parameterizing your manifests, and loading & templating in a single command.

#
# Given the following YAML template
#
apiVersion: v1
kind: Pod
metadata:
  name: nginx
  labels:
    name: nginx
spec:
  containers:
  - name: nginx
    image: ${docker_image}
    ports:
    - containerPort: 80


#
# Load the yaml file, parsing the ${docker_image} variable
#
data "kubectl_path_documents" "manifests" {
    pattern = "./manifests/*.yaml"
    vars = {
        docker_image = "https://myregistry.example.com/nginx"
    }
}

All the terraform hcl template features are supported (e.g. directives).

You can now also disable the waiting for rollout behavior via the wait_for_rollout parameter. By default, the kubectl-provider will wait for Deployments/DaemonSets/APIServices to finish rollout before proceeding. Set this flag to false to disable this behavior.

#10 - Support for exec provider configuration added - 🎩 @rinrailin

Documentation on configuring available at: https://gavinbunney.github.io/terraform-provider-kubectl/docs/provider.html#exec-plugin-support

PR #3 - Fix issues with large yaml files splitting over default scanner buffer limit