Skip to main content
Version: Unreleased ๐Ÿšง

Generating manifests

This document describes how to generate Manifests from existing Helm Charts and Terraform Modules. In this tutorial you will see how to generate an Interface for the Redis database. You will generate Implementations for the Bitnami Helm Chart and AWS Terraform module.


  • Read Content development guide before.
  • git installed.
  • Helm v3 installed.
  • Capact CLI at least v0.5.0 installed.

    NOTE: Install the latest Capact CLI version from the main branch.

  • Capact cluster. For example, local instance.

    NOTE: Use --capact-overrides=hub-public.populator.enabled=false flag, as you will manually upload your OCF manifests into Hub.

Generating Interface#

To create an Interface for the Redis database, first you need to choose a path for it. We already have a path cap.interface.database with PostgreSQL and MongoDB there, so let's use cap.interface.database.redis.install:

capact alpha manifest-gen interface cap.interface.database.redis.install

It generates four files:

  • generated/interface/database/redis.yaml with an InterfaceGroup definition. As the name suggests, it's grouping interfaces for Redis domain. For example, install, update etc.
  • generated/interface/database/redis/install.yaml with an Interface definition. It defines the Interface signature.
  • generated/type/database/redis/install-input.yaml with an input Type definition.
  • generated/type/database/redis/config.yaml with an output Type definition.

All the files have a metadata property defined with fields like description, maintainer, name, e-mail and so on. Feel free to update them now. For example, metadata in redis.yaml file can look like this:

metadata:  prefix: "cap.interface.database.redis"  name: "install"  displayName: "Install Redis"  description: "Install action for the Redis database"  documentationURL:  supportURL:  iconURL:  maintainers:    - email: team-[email protected]      name: Capact Dev Team      url:

Adjusting schema#

By default, Type install-input is empty. To be able to pass an input, you need to adjust schema to accept a cluster name. Edit file generated/type/database/redis/install-input.yaml and replace empty schema with:

    value: |-      {        "$schema": "",        "type": "object",        "title": "The schema for Redis installation input",        "examples": [          {            "name": "Cluster name",            "password": "foo"          }        ],        "required": [           "name"        ],        "properties": {          "name": {            "$id": "#/properties/name",            "type": "string",            "title": "Defines a cluster name for the Redis"          },          "password": {            "$id": "#/properties/password",            "type": "string",            "title": "Defines a password"          }        },        "additionalProperties": false      }

Edit file generated/type/database/redis/config.yaml and replace empty schema with:

    value: |-      {        "$schema": "",        "type": "object",        "title": "The schema for Redis configuration",        "examples": [          {            "password": "pass"          }        ],        "required": [          "host",          "port",          "password"        ],        "definitions": {          "hostname": {            "type": "string",            "format": "hostname",            "title": "Hostname"          },          "port": {            "type": "integer",            "title": "Port",            "minimum": 0,            "maximum": 65535          }        },        "properties": {          "password": {            "$id": "#/properties/rootPassword",            "type": "string",            "title": "Defines a database password"          },          "host": {            "$ref": "#/definitions/hostname"          },          "port": {            "$ref": "#/definitions/port"          }        },        "additionalProperties": false      }

Generating Helm Implementation#

Next step is to generate Implementation for a Helm Chart. First step is to clone git repository with Bitnami Charts:

helm repo add bitnami pull bitnami/redis --untar --untardir charts/bitnami

Now you can generate a new Implementation manifest. Set a path to the downloaded chart and specify which Interface is implemented:

capact alpha manifest-gen implementation helm cap.implementation.bitnami.redis.install charts/bitnami/redis --repo "" -i cap.interface.database.redis.install

It generates two files:

  • generated/type/bitnami/redis/install-input.yaml with the Type for Implementation specific additional input.
  • generated/implementation/bitnami/redis/install.yaml with the Implementation definition for the Redis chart.

Similarly to the Interface, you can update metadata in both files. Now you need to adjust manifests to use input values from the Interface. Modify files according to the following diffs:

--- a/generated/implementation/bitnami/redis/install.yaml+++ b/generated/implementation/bitnami/redis/install.yaml@@ -104,7 +104,7 @@ spec:                                 enabled: <@ additionalinput.auth.enabled | default(true) | tojson @>                                 existingSecret: <@ additionalinput.auth.existingSecret | default("") | tojson @>                                 existingSecretPasswordKey: <@ additionalinput.auth.existingSecretPasswordKey | default("") | tojson @>-                                password: <@ additionalinput.auth.password | default("") | tojson @>+                                password: <@ input.password | default(random_word(length=16)) @>                                 sentinel: <@ additionalinput.auth.sentinel | default(true) | tojson @>                                 usePasswordFiles: <@ additionalinput.auth.usePasswordFiles | default(false) | tojson @>                               clusterDomain: <@ additionalinput.clusterDomain | default("cluster.local") @>@@ -121,7 +121,7 @@ spec:                                 enabled: <@ additionalinput.diagnosticMode.enabled | default(false) | tojson @>                               existingConfigmap: <@ additionalinput.existingConfigmap | default("") | tojson @>                               extraDeploy: <@ additionalinput.extraDeploy | default(None) | tojson @>-                              fullnameOverride: <@ additionalinput.fullnameOverride | default("") | tojson @>+                              fullnameOverride: <@ @>                               global:                                 imagePullSecrets: <@ | default(None) | tojson @>                                 imageRegistry: <@ | default("") | tojson @>@@ -467,7 +467,9 @@ spec:
                             output:                               goTemplate: |-                                # TODO(ContentDeveloper): Add output template in YAML+                                host: '{{ template "common.names.fullname" . }}.{{ .Release.Namespace }}'+                                port: '{{ .Values.master.service.port }}'+                                password: '{{ .Values.auth.password }}'               - - name: helm-install                   capact-action: helm.install
--- a/generated/type/bitnami/redis/install-input.yaml+++ b/generated/type/bitnami/redis/install-input.yaml@@ -36,16 +36,6 @@ spec:                 "type": "boolean",                 "form": true,                 "title": "Use password authentication"-              },-              "password": {-                "type": "string",-                "title": "Redis password",-                "form": true,-                "description": "Defaults to a random 10-character alphanumeric string if not set",-                "hidden": {-                  "value": false,-                  "path": "auth/enabled"-                }               }             }           },

Now your Helm manifests are ready to use! See how can you test it.

Generating Terraform Implementation#

Next step is to generate Implementation for a Terraform module. First, clone git repository with the module:

git clone --depth 1

Now you can generate a new Implementation manifest. You need to set a path to the module and set which Interface is implemented:

capact alpha manifest-gen implementation terraform terraform-aws-elasticache-redis -i cap.interface.database.redis.install -s git:: -p aws

It generates two new files:

  • generated/type/aws/redis/install-input.yaml with the Type for Implementation-specific additional input.
  • generated/implementation/aws/redis/install.yaml with the Implementation definition for AWS ElastiCache Terraform module.

Similarly to the Interface, you can update metadata in both files.

Now you need to adjust manifests to use input values from the Interface. Modify files according to the following diffs:

--- a/generated/type/aws/redis/install-input.yaml+++ b/generated/type/aws/redis/install-input.yaml@@ -14,10 +14,13 @@ metadata:       url: spec:   jsonSchema:-    # TODO(ContentDeveloper): Adjust the JSON schema if needed.     value: |-       {         "properties": {+          "region": {+            "type": "string",+            "title": "AWS region"+          },           "apply_immediately": {             "type": "boolean",             "title": "apply_immediately",@@ -28,11 +31,6 @@ spec:             "title": "at_rest_encryption_enabled",             "description": "Whether to enable encryption at rest."           },-          "auth_token": {-            "type": "string",-            "title": "auth_token",-            "description": "The password used to access a password protected server. Can be specified only if `transit_encryption_enabled = true`."-          },           "auto_minor_version_upgrade": {             "type": "string",             "title": "auto_minor_version_upgrade"@@ -102,11 +100,6 @@ spec:             "title": "multi_az_enabled",             "description": "Specifies whether to enable Multi-AZ Support for the replication group. If true, `automatic_failover_enabled` must also be enabled. Defaults to false."           },-          "name_prefix": {-            "type": "string",-            "title": "name_prefix",-            "description": "The replication group identifier. This parameter is stored as a lowercase string."-          },           "node_type": {             "type": "string",             "title": "node_type",
--- a/generated/implementation/aws/redis/install.yaml+++ b/generated/implementation/aws/redis/install.yaml@@ -16,7 +16,7 @@ metadata:     name: "Apache 2.0" spec:-  appVersion: "1.0.x" # TODO(ContentDeveloper): Set the supported application version here+  appVersion: "6.2.x"   additionalInput:     parameters:       additional-parameters:@@ -104,6 +104,7 @@ spec:                             env:                               - AWS_ACCESS_KEY_ID=<@ providercredentials.accessKeyID @>                               - AWS_SECRET_ACCESS_KEY=<@ providercredentials.secretAccessKey @>+                              - AWS_DEFAULT_REGION=<@ additionalinput.region | default('eu-west-1') @>                             output:                               goTemplate:                                 elasticache_auth_token: "{{ .elasticache_auth_token }}"@@ -123,15 +124,15 @@ spec:                                 security_group_owner_id: "{{ .security_group_owner_id }}"                                 security_group_vpc_id: "{{ .security_group_vpc_id }}"                             variables: |++                              name_prefix = <@ | tojson @>+                              auth_token = <@ input.password | default(random_word(length=16)) | tojson @>+                               <%- if additionalinput.apply_immediately %>                               apply_immediately = <@ additionalinput.apply_immediately | tojson @>                               <%- endif %>                               <%- if additionalinput.at_rest_encryption_enabled %>                               at_rest_encryption_enabled = <@ additionalinput.at_rest_encryption_enabled | tojson @>                               <%- endif %>-                              <%- if additionalinput.auth_token %>-                              auth_token = <@ additionalinput.auth_token | tojson @>-                              <%- endif %>                               <%- if additionalinput.auto_minor_version_upgrade %>                               auto_minor_version_upgrade = <@ additionalinput.auto_minor_version_upgrade | tojson @>                               <%- endif %>@@ -174,9 +172,6 @@ spec:                               <%- if additionalinput.multi_az_enabled %>                               multi_az_enabled = <@ additionalinput.multi_az_enabled | tojson @>                               <%- endif %>-                              <%- if additionalinput.name_prefix %>-                              name_prefix = <@ additionalinput.name_prefix | tojson @>-                              <%- endif %>                               <%- if additionalinput.node_type %>                               node_type = <@ additionalinput.node_type | tojson @>                               <%- endif %>@@ -246,9 +241,10 @@ spec:                           data: ""                       - name: template                         raw:-                          # TODO(ContentDeveloper): Fill the properties of the output TypeInstance here                           data: |-                            property: value+                            password: "<@ elasticache_auth_token @>"+                            port: "<@ elasticache_port @>"+                            host: "<@ elasticache_replication_group_primary_endpoint_address @>"           - name: prepare-parameters             inputs:

Now your Terraform manifests are ready to use! See how can you test it.

Testing manifests#

To test generated manifests you need to:

  1. Clone repository with Hub Manifests.

  2. Copy content of the generated directory into manifests directory.

  3. Populate the manifests to Public Hub.

  4. (For AWS only) Create a Type Instance with credentials.

  5. Prepare input file input.yaml

     cat > /tmp/install-input.yaml << ENDOFFILE input-parameters:   name: redis-test  ENDOFFILE
  6. Create an action

    • For Helm Implementation

      capact act create cap.interface.database.redis.install --name redis --parameters-from-file /tmp/install-input.yaml
    • To choose Terraform implementation create a policy file policy.yaml

      AWS_TI_ID="{AWS Type Instance ID}"AWS_VPC_ID="{Your VPC ID}"AWS_SUBNET_ID="{Your VPC subnet id}"cat > /tmp/aws-policy.yaml << ENDOFFILErules:  - interface:      path: cap.interface.database.redis.install    oneOf:    - implementationConstraints:        path: ""      inject:        requiredTypeInstances:          - id: "${AWS_TI_ID}"            description: "AWS credentials"        additionalParameters:          - name: additional-parameters            value:              vpc_id: "${AWS_VPC_ID}"              subnet_ids:                - "${AWS_SUBNET_ID}"              number_cache_clusters: 1              node_type: cache.t3.smallENDOFFILE

      Create an action

      capact act create cap.interface.database.redis.install --name redis --parameters-from-file /tmp/install-input.yaml --action-policy-from-file /tmp/aws-policy.yaml
  7. Get the status of the Action from the previous step:

    capact act get redis

    In the STATUS column you can see the current status of the Action. When the Action workflow is being rendered by the Engine, you will see the BEING_RENDERED status. After the Action finished rendering and the status is READY_TO_RUN, you can go to the next step.

  8. Run the rendered Action

    After the Action is in READY_TO_RUN status, you can run it. To do this, execute the following command:

    capact act run redis
  9. Check the Action execution and wait till it is finished:

    capact action watch redis
  10. Get the ID of the cap.type.database.redis.config TypeInstance:

    capact action get redis -ojson | jq -r '.Actions[].output.typeInstances | map(select(.typeRef.path == "cap.type.database.redis.config"))'
  11. Get the TypeInstance value:

    Use the ID from the previous step and fetch the TypeInstance value:

    capact typeinstance get {type-instance-id} -ojson | jq -r '.[0].latestResourceVersion.spec.value'
  12. Use the information from the TypeInstance to connect to the Redis.


That's all. Now you can share you generated manifests. Create a pull request in our Hub Manifests repository. We are happy to review it :)