🧬 helmwave.yml¶

This documentation describes all entities inside a helmwave.yml

project: "HelloWorld"
version: "0.38.0"
repositories: []
registries: []
releases: []
lifecycle:
  pre_up: []
  post_up: []
  pre_down: []
  post_down: []
  pre_build: []
  post_build: []
  pre_rollback: []
  post_rollback: []

project: "HelloWorld"
version: "0.38.0"

repositories:
    - name: stable
      url: https://localhost:8080
      # private repositories
      username: user
      password: 12345679

      # cert
      certFile: ./cert.pem
      keyFile: ./key.pem
      caFile: ./ca.pem

lifecycle:
    pre_build:
      - cmd: "ls"
        args: ["-l", "-a"]
        show: false
        allow_failure: false
      - echo "run global pre_build script"
    post_build:
      - echo "run global post_build script"

registries:
    - host: https://localhost:5000
      # private registries
      username: oci_user
      password: 98765431
      insecure: false


releases:
    - name: my
      namespace: prod
      chart:
        name: stable/my-chart
        version: "0.2.0"
        ca_file: ""
        cert_file: ""
        key_file: ""
        insecure: false
        keyring: ""
        password: ""
        pass_credentials: false
        repo_url: ""
        username: ""
        verify: false

      store:
        greeting: "HelloWorld"

      lifecycle:
        pre_up:
          - echo "running pre_up script for my"
        post_up:
          - echo "running post_up script for my"
        pre_build:
          - echo "running pre_build script for my"
        post_build:
          - echo "running post_build script for my"

      depends_on:
        - name: db@prod
          tag: prod
          optional: false

      values:
        - src: values.yml
          delimiter_left: "[["
          delimiter_right: "]]"
          render: false
          strict: true

      tags:
        - my
        - product

      post_renderer: "./gomplate"
      offline_kube_version: "1.22.0"
      timeout: "5m"
      max_history: 3
      context: prod
      description: "my chart"
      pending_release_strategy: rollback
      allow_failure: false
      atomic: true
      cleanup_on_fail: false
      create_namespace: false
      disable_hooks: false
      disable_openapi_validation: false
      force: false
      enable_dns: false
      recreate: false
      reset_values: false
      reuse_values: false
      skip_crds: false
      sub_notes: false
      wait: true
      wait_for_jobs: true

project¶

Reserved for the future.

version¶

This version for helmwave.

Helmwave will check the current version and project version.

In the future, it is planned to check major compatibility.

Note that this is the version of the Helmwave itself (not your service/application) to check for compatibility of the schema, version, etc.

RECOMMENDED: Set this explicitly to helmwave version command result to be safe later on.

registries[]¶

Introduced in v0.19.0

Describe which OCI registries need to be added.

project: "Example: oci private"
version: "0.38.0"

registries:
  - host: localhost:5000
    username: <pass_username_here>
    password: <pass_password_here>

releases:
  - name: my
    chart:
      name: oci://localhost:5000/my-chart
      version: 0.2.0 # best practice

project: "Example: oci public"
version: "0.38.0"

registries:
  - host: ghcr.io

releases:
  - name: my
    chart:
      name: oci://ghcr.io/helmwave/unit-test-oci
      version: 0.1.0 # best practice

host¶

Host of the registry.

username¶

Username for the registry.

password¶

Password for the registry.

repositories[]¶

helm repositories also know as helm repo add

This repository will be stored in a local helm repositories database.

name¶

Local name alias.

url¶

URL of the repository.

username¶

Username for the repository.

password¶

Password for the repository.

force¶

Update existing repository exists if settings differ.

lifecycle¶

Aka global hooks. Introduced in v0.28.0

We don't call lifecycle the hooks on purpose so as not to confuse you with the original functionality of helm hooks.

environment

Introduced in v0.29.0

To each lifecycle command several environment variables are passed:

• ${HELMWAVE_LIFECYCLE_TYPE} - contains lifecycle stage/type (pre_build/post_build/etc.)
• ${HELMWAVE_LIFECYCLE_RELEASE_UNIQNAME} - (only for per-release lifecycle) contains release uniqname (release@namespace)

Hook

example

cmd, args¶

version: 0.38.0
lifecycle:
  pre_build:
    - echo "run global pre_build script"

version: 0.38.0
lifecycle:
  pre_build:
    - cmd: echo
      args:
        - "run global pre_build script"
      show: true
      allow_failure: false

show¶

Introduced in v0.28.0

Show output of the command.

allow_failure¶

Introduced in v0.29.0

Allow failure of the command.

monitors[]¶

Introduced in v0.32.0

Monitors run after all dependant releases succeeded. They are triggered every interval and end when:

• Either success_threshold is met - monitor succeeds
• Either failure_threshold is met - monitor fails
• Or total_timeout is triggered - monitor fails

example

name¶

Monitor name, must be unique.

type¶

Type of monitor to run. Must be one of:

• prometheus
• http

total_timeout¶

Total timeout for whole monitor run. If monitor haven't finished within total_timeout, it is considered as failed.

iteration_timout¶

Timeout for each monitor iteration. If monitor iteration haven't finished within iteration_timeout, iteration is considered as failed.

interval¶

Interval between iterations. Must be lower than iteration_timeout.

success_threshold¶

Count of sequential succeeded iterations that make monitor succeed.

failure_threshold¶

Count of sequential failed iterations that make monitor fail.

prometheus¶

url¶

Prometheus base URL

expr¶

Prometheus expression to query on each iteration. Iteration is considered successful if expression returns 1 or more rows.

insecure¶

Whether to skip SSL certificate validation

http¶

url¶

URL to query

expected_codes¶

List of expected HTTP response codes that are considered to be successful

method¶

HTTP method to query

body¶

HTTP body to send

headers¶

Map of HTTP headers to set in request

insecure¶

Whether to skip SSL certificate validation

releases[]¶

Almost all options that are here are native helm options.

name¶

Introduced in v0.5.0

Release name. I hope you know what it is.

namespace¶

Introduced in v0.5.0

kubernetes namespace.

chart¶

Introduced in v0.5.0

chart can be an object or a string. If it's a string, it will be treated as a name.

releases:
  - name: my-release
    namespace: my-namespace
    chart: my-chart

releases:
  - name: my-release
    namespace: my-namespace
    chart:
      name: my-chart

!!! tip "If chart is remote it will be downloaded into .helmwave/charts and downloaded archive will be used during deploy."

name¶

Can be oci://, local or repository chart.

releases:
  - name: my-release
    namespace: my-namespace
    chart: oci://my-registry.io/my-chart

releases:
  - name: my-release
    namespace: my-namespace
    chart: ./my-chart

releases:
  - name: my-release
    namespace: my-namespace
    chart: bitnami/nats

version¶

Chart version. If not set, latest will be used.

releases:
  - name: my-release
    namespace: my-namespace
    chart:
      name: my-chart
      version: 1.2.3

skip_dependency_update¶

Introduced in v0.29.1

Disable Helm dependency update.

skip_refresh¶

Introduced in v0.29.1

Disable Helm repository refresh.

create_namespace¶

Introduced in v0.12.0

If set to true Helmwave will create the release namespace if not present.

helm upgrade --install --create-namespace my-release my-chart --namespace my-namespace

releases:
  - name: my-release
    namespace: my-namespace
    chart: my-chart
    create_namespace: true

values[]¶

Introduced in v0.5.0

values can be an object or a string. If it's a string, it will be treated as a src field.

values:
  - values.yaml
  - values2.yaml

values:
  - src: values.yaml
    delimiter_left: "{{"
    delimiter_right: "}}"
    strict: false
    renderer: ""

src¶

Path to values file. It can be local or remote.

delimiter_left, delimiter_right¶

Introduced in v0.24.0

You can change the delimiter that helmwave uses to render values.

example

renderer¶

Introduced in v0.32.0

Allows overriding how values file is rendered. These renderers are supported:

• sprig - template file with sprig (overrides global --templater argument)
• gomplate - template file with gomplate (overrides global --templater argument)
• copy - do not template file at all (overrides global --templater argument)

• sops - use sops to decode encrypted file

• example

• example

strict¶

Introduced in v0.20.0

Allows to fail if values file doesn't exist.

example

tags[]¶

Aka labels. Introduced in v0.4.0

Tags allow you to choose releases for build.

example

offline_kube_version¶

Introduced in v0.27.3

If offline_kube_version set helmwave will use this version to build plan. Without this option, helmwave will ask kubernetes for a version. It is very useful if you want to build a plan without access to a cluster.

Combine offline_kube_version and --diff-mode=local or --diff-mode=none to build a plan without kubernetes.

example

store¶

Introduced in v0.2.0

It allows passing your custom fields from helmwave.yml to values.

example

lifecycle¶

Aka hooks. Introduced in v0.28.0

example

depends_on[]¶

Introduced in v0.9.0

depends_on can be an object or a string. If it's a string, it will be treated as a name.

depends_on is a list of releases that allow you to deploy a sequence.

!!! example "Example for 3-tier application"

```mermaid
graph LR
    frontend --> backend --> db;
```
*If you don't see a graph, please reload the page.*

Your helmwave.yml should look like this:

releases:
  - name: frontend
    depends_on:
      - backend
    namespace: test

  - name: backend
    depends_on:
      - db
    namespace: test

  - name: db
    allow_failure: false
    namespace: test

name¶

Name of release (dependency) that has to be installed/upgraded before this release (dependant). If dependency is not in a plan, it will be added to a plan.

Name support 2 kind of definitions: uniq name <release-name>@<namespace> or just <release-name>. If namespace is not specified, it will be taken from namespace filed of release.

The same configuration can be written in 2 ways:

releases:
  - name: backend
    namespace: test
    depends_on:
      - name: redis

- name: redis
  namespace: test

releases:
  - name: backend
    namespace: test
    depends_on:
      - name: redis@test

- name: redis
  namespace: test

Both of them will be normalized to redis@test in a planfile.

tag¶

Introduced in v0.24.0

You can include all releases that match this tag to be added as dependencies. If a tag is not in a plan, it will be added to a plan.

The planfile (.helmwave/planfile by default) will have a normalized list of releases instead of tags.

optional¶

Introduced in v0.24.0

If dependency is not found in all available releases, helmwave will not fail due to missing dependency.

It allows setting explicit dependencies between releases. Dependant release will start upgrading only after all its dependencies finished upgrading

monitors[]¶

Introduced in v0.32.0

Which monitors should be triggered after the releases is successfully deployed.

example

name¶

Name of the monitor to trigger

action¶

Action to perform if the monitor is failed. Should be one of:

• "" - do nothing
• rollback - rollback release to the previous version
• uninstall - uninstall release

If multiple monitors of the release failed and they have different actions, chosen action is a random one from the list.

allow_failure¶

Introduced in v0.15.0

Allows all dependant releases to proceed even if release failed.

pending_release_strategy¶

Introduced in v0.21.0

Strategy to handle releases in pending statuses (pending-install, pending-upgrade, pending-rollback)

If helmwave tries to upgrade release that is currently in one of the pending statuses, it will follow specified strategy:

• "" (or not specified) - do nothing. helm will fail in this case;
• rollback - rollback release to a previous version. Upgrade will happen after rollback is complete;
• uninstall - uninstall release. Upgrade will happen after uninstall is complete.

wait¶

We recommend using wait for all releases. It will wait for all resources to be ready.

wait_for_jobs¶

Introduced in v0.12.0

We recommend using wait_for_jobs for all releases. It will wait for all jobs to be completed.

timeout¶

Introduced in v0.5.0

Time to wait for release to install.

max_history¶

Introduced in v0.5.0

Limit the maximum number of revisions saved per release. Use 0 for no limit (default 0)

context¶

Introduced in v0.24.0

Allows using custom kube-context for release.

description¶

Introduced in v0.12.0

Additional information about release.

atomic¶

Introduced in v0.5.0

if set, upgrade process rolls back changes made in case of failed upgrade.

cleanup_on_fail¶

Introduced in v0.12.0

Allow deletion of new resources created in this upgrade when upgrade fails.

devel¶

Removed in v0.29.3 as it wasn't functioning properly.

If you need to use development version, set version: ">0.0.0-0".

disable_hooks¶

Introduced in v0.12.0

Disable pre/post upgrade helm hooks

disable_open_api_validation¶

Introduced in v0.12.0

If set, the upgrade process will not validate rendered templates against the Kubernetes OpenAPI Schema

force¶

Introduced in v0.5.0

Force resource updates through a replacement strategy

enable_dns¶

Introduced in v0.27.1

Enable DNS resolution in templates.

recreate¶

Introduced in v0.5.0

Allows deleting and then creating resources (pods) when needed instead of updating.

reset_values¶

Introduced in v0.5.0

When upgrading, reset the values to the ones built into the chart.

reuse_values¶

Introduced in v0.5.0

When upgrading, reuse the last release's values and merge it into the new configuration.

reset_then_reuse_values¶

Introduced in v0.36.0

When upgrading, reset the values to the ones built into the chart, apply the last release's values and merge in any overrides from the values files. If 'reset_values' or 'reuse_values' is specified, this is ignored.

skip_crds¶

Introduced in v0.12.0

If set, no CRDs will be installed. By default, CRDs are installed if not already present.

Be careful with this option.

• If you have a release that depends on CRDs, it will fail.
• Sometimes you should enable installCRD: true in your chart like cert-manager.

sub_notes¶

Introduced in v0.12.0

If set, render sub chart notes along with the parent. This option determines whether sub-notes are rendered in the chart.

post_renderer¶

Introduced in v0.24.0

You can use custom commands to change rendered manifests.

show_notes¶

Introduced in v0.31.0

If enabled, rendered chart notes will be shown after successful release.

tests¶

Introduced in v0.34.0

enabled¶

Whether to run helm chart tests.

force_show_logs¶

If enabled, helmwave will show logs even if tests are successful.

filters¶

Map of filters to run only specific tests.

deletion_propagation¶

Introduced in v0.36.3

Sets deletion cascading strategy. Should be one of background, foreground, orphan.