feat(rfc): new artifacts service

.gitignore

@@ -1 +1,3 @@
 .idea
+
+.DS_Store

rfc/new-artifacts-service/README.md

@@ -0,0 +1,207 @@
+# Separate artifacts to new service
+
+| | |
+|-|-|
+| **Status**     | **Proposed**, Accepted, Implemented, Obsolete |
+| **RFC #**      | _Update after PR has been made_ |
+| **Author(s)**  | [Nemesis Osorio](https://github.com/nemesisOsorio), [Sergio Quintero](https://github.com/sergio-quintero) |
+| **SIG / WG**   | _Applicable SIG(s) or Working Group(s)_ |
+
+## Overview
+
+Users with large context pipelines do not have a good experience with pipeline executions. <br/>
+Context and outputs in pipeline executions can contain large artifacts/objects such as k8s manifest (yaml/json files). <br/>
+Spinnaker passes these artifacts/objects in request and response bodies across some services. <br/>
+The idea is to create a new service to store these artifacts/objects
+and only pass the `id`s of those artifacts/objects in the request/response bodies and context/outputs.
+
+Deck users do not click to see the produced artifact/object in every single pipeline execution.
+With this approach, artifacts will be loaded on demand (lazy loading).
+
+### Goals and Non-Goals
+
+#### Goals
+
+* Improve Deck performance with large pipeline executions.
+* Optimize message size over the wire.
+
+#### Non-Goals
+
+* Move [clouddriver-artifacts](https://github.com/spinnaker/clouddriver/blob/master/clouddriver-artifacts)
+to the new service
+* Refactor Angular code in Deck.
+* Add new permissions to Fiat service.
+* Avoid the duplication of artifacts when executing a pipeline multiple times without changes.
+
+## Motivation and Rationale
+
+* Deck is very slow and sometimes freezes when you are producing/consuming large artifacts/objects in pipeline executions.
+* Smaller payloads are sent in request/response.
+* Save serializing and deserializing computations.
+
+Early customers:
+* Armory on behalf of its customers.
+
+## Timeline
+
+The proposed plan is to implement this by pipeline stages:
+
+* Milestone 1: (Targeting 1.32) Address stages with the most acute performance issues
+  * Introduce new Artifact service
+  * Bake (Manifest), both Helm and Kustomize providers
+  * Deploy (Manifest)
+* Milestone 2: Analyze additional stages that exhibit similar performance patterns
+* Milestone 3: (Targeting 1.33) Integrate additional stages with artifact service
+* Milestone 4: Support to deploy artifacts service in Operator
+
+## Design
+
+Creation of a new microservice for artifacts/objects. <br/>
+We are proposing the following API: [Artifacts API](./artifacts-api.yml), you can render it using [Swagger Editor](https://editor-next.swagger.io/) <br/>
+Support of the following artifact types:
+* base64
+* json (everything in the execution details has json format)
+
+```mermaid
+---
+title: ER Diagram
+---
+erDiagram
+    artifacts {
+        uuid id PK "artifact id"
+        varchar name "artifact name"
+        varchar type "base64, json"
+        json content "the large artifact"
+        timestamp created_at "creation date time"
+    }
+```
+
+Make changes in Deck components that are using the context/output to render artifacts/objects,
+and only load those artifacts/objects if you click the button/link(lazy loading).
+
+Enable Fiat autoconfiguration and use application permissions for new artifacts endpoints.
+
+### Retention Policy
+
+Add a scheduler to automatically delete artifacts. The following options will be configurable:
+* Retention time: How long the record will be kept in the database.
+* Artifact cleanup job: Configurable cron job that deletes artifacts periodically in accordance to the retention time property.
+
+### Dependencies
+
+* A new database.
+  * Support for MySQL and PostgreSQL.
+  * Database changelogs managed by liquibase.
+* Services involved
+  * Orca
+  * Clouddriver
+  * Rosco
+  * Deck
+  * Gate
+
+All changes will be implemented under a feature flag.
+
+```mermaid
+---
+title: Artifacts Interactions
+---
+flowchart TB
+    subgraph Artifacts
+        direction TB
+        Rest-Service --> DB
+        Scheduler --> DB
+        DB[(DB)]
+    end
+    Orca --> Artifacts
+    Clouddriver --> Artifacts
+    Rosco --> Artifacts
+    Gate --> Artifacts
+    Deck --> Gate
+    Artifacts --> Fiat
+```
+
+### Bake (Manifest) example
+
+![Bake Pipeline](./bake-pipeline.png "Bake Manifest Pipeline")
+
+```mermaid
+sequenceDiagram
+title Bake Manifest
+    participant Deck
+    participant Gate
+    participant Orca
+    participant Rosco
+    participant Artifacts
+    Deck->>+Gate: startPipelineExecution
+    Gate->>+Orca: startPipelineExecution
+    Orca-->>-Gate: #nbsp;
+    Gate-->>-Deck: #nbsp;
+    Orca->>+Rosco: bakeManifest
+    Rosco->>+Artifacts: createArtifact
+    Artifacts-->>-Rosco: artifactDetails
+    Rosco-->>-Orca: artifactDetails
+```
+
+```mermaid
+sequenceDiagram
+title Get Baked Manifest YAML
+    participant Deck
+    participant Gate
+    participant Artifacts
+    Deck->>+Gate: Baked Manifest YAML
+    Gate->>+Artifacts: getArtifactById
+    Artifacts-->>-Gate: artifact
+    Gate-->>-Deck: artifact
+```
+
+Diagrams: you can render them using [Mermaid Live Editor](https://mermaid.live/)
+or the cli `mmdc -i README.md`
+
+## Drawbacks
+
+* This introduces a new service that adds complexity.
+* Implementing it under a feature flag means we need to ensure it does not break existing users.
+
+## Alternatives
+
+An example of how it could be implemented in Orca without introducing a new service:
+
+1. Orca will request a bake operation to Rosco.
+   * Rosco performs a bake and returns the artifact in base64.
+   * Orca will receive the response and store the artifact.
+   * Orca will **only** add the artifactId to the pipeline context/outputs.
+2. When deploying a Manifest
+   * Orca will fetch the artifact by id and add the artifact to the request.
+   * Clouddriver will continue to receive the manifest in the request body.
+
+Notes:
+* Only Orca would change the behavior, other services remain the same.
+* Add the new table to Orca's database schema.
+
+## Known Unknowns
+
+* Not all the stages/services that can produce/consume those artifacts/objects are known.
+* How will artifacts be passed in automated triggers?
+* How to migrate/support existing executions?
+
+## Security, Privacy, and Compliance
+
+* Enable Fiat autoconfiguration and use application permissions
+
+## Operations
+
+* The new service will require a new repository and integration to the Spinnaker CI/CD process.
+* The database maintenance plan.
+
+## Risks
+
+A scenario when a user enables the feature flag to use the new artifacts service,
+runs a pipeline with artifacts and later disables the feature, leaving orphaned artifacts.
+
+## Future Possibilities
+
+* Allow saving multiple artifacts at once
+* Add other implementations for the persistence layer
+  * s3
+* Move [clouddriver-artifacts](https://github.com/spinnaker/clouddriver/blob/master/clouddriver-artifacts)
+to the new artifacts service

rfc/new-artifacts-service/artifacts-api.yml

@@ -0,0 +1,113 @@
+openapi: 3.0.3
+info:
+  title: Artifacts
+  description: |-
+    Artifacts or objects that could be large in pipeline executions
+  version: 0.0.0
+servers:
+  - url: https://items.spinnaker.io/api/v0
+tags:
+  - name: artifacts
+    description: Everything about your Artifacts
+paths:
+  /artifacts:
+    post:
+      tags:
+        - artifacts
+      summary: Add a new artifact
+      description: Add a new artifact
+      operationId: addArtifact
+      requestBody:
+        description: Create a new Artifact
+        content:
+          application/json:
+            schema:
+              $ref: '#/components/schemas/Artifact'
+        required: true
+      responses:
+        '201':
+          description: Artifact created
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/Artifact'
+  /artifacts/{artifactId}:
+    get:
+      tags:
+        - artifacts
+      summary: Find Artifact by id
+      description: Returns a single artifact
+      operationId: getArtifactById
+      parameters:
+        - name: artifactId
+          in: path
+          description: Id of artifact to return
+          required: true
+          schema:
+            type: integer
+            format: int64
+      responses:
+        '200':
+          description: successful operation
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/Artifact'
+        '404':
+          description: Artifact not found
+    delete:
+      tags:
+        - artifacts
+      summary: Deletes an Artifact
+      description: delete an artifact
+      operationId: deleteArtifact
+      parameters:
+        - name: artifactId
+          in: path
+          description: Artifact id to delete
+          required: true
+          schema:
+            type: integer
+            format: int64
+      responses:
+        '204':
+          description: successful operation
+components:
+  schemas:
+    Artifact:
+      required:
+        - name
+        - type
+      type: object
+      properties:
+        id:
+          type: string
+          format: uuid
+        name:
+          type: string
+          example: kustomize
+        content:
+          $ref: '#/components/schemas/Content'
+        type:
+          type: string
+          description: artifact type
+          enum:
+            - base64
+            - json
+        createdAt:
+          type: string
+          format: date-time
+    Content:
+      description: content of the artifact
+      anyOf:
+        - type: string
+        - type: array
+          items: {}
+        - type: object
+  requestBodies:
+    Artifact:
+      description: Artifact
+      content:
+        application/json:
+          schema:
+            $ref: '#/components/schemas/Artifact'