Backfill workspace_agentk_states for all workspaces

Issue: Add a background DB migration to populate the w... (#538360 - closed)

What does this MR do and why?

Backfill workspace_agentk_states for all workspaces

This migration will create desired_config for all the workspaces regardless of their state and create an entry in the table workspace_agentk_states. This is required as part of freezing desired_config of Workspaces in the create step instead of reconciliation.

Changelog: other

EE: true

References

Some of the related MRs towards this efforts are

1. Add files from RemoteDevelopment module for mig... (!199244 - merged)
2. Validate only kubernetes objects using JSON schema (!200985 - merged)
3. Update the duplicated logic to create desired c... (!202190 - merged)
4. Remove bm_desired_config_array_validator_spec (!202551 - merged)

How to set up and validate locally

1. Run the migration using bin/rails db:migrate
2. Notice the table workspace_agentk_states has rows corresponding to each workspace.

For database reviewers

Query

INSERT INTO "workspaces" ("created_at", "updated_at", "user_id", "project_id", "cluster_agent_id", "desired_state_updated_at", "responded_to_agent_at", "name", "namespace", "desired_state", "actual_state", "devfile_path", "devfile", "processed_devfile", "url", "deployment_resource_version", "personal_access_token_id", "workspaces_agent_config_version", "desired_config_generator_version", "project_ref", "actual_state_updated_at") VALUES ('2025-09-03 13:08:01.072419', '2025-09-03 13:08:01.072419', 60, 60, 1, '2025-09-03 13:08:01.069925', '2025-09-03 13:08:01.069925', 'workspace-2', 'workspace_2_namespace', 'Running', 'Running', 'devfile-path', 'schemaVersion: 2.2.0
components:
  - name: tooling-container
    attributes:
      gl/inject-editor: true
    container:
      image: registry.gitlab.com/gitlab-org/remote-development/gitlab-remote-development-docs/debian-bullseye-ruby-3.2-node-18.12:rubygems-3.4-git-2.33-lfs-2.9-yarn-1.22-graphicsmagick-1.3.36-gitlab-workspaces
      env:
        - name: KEY
          value: VALUE
      endpoints:
      - name: http-3000
        targetPort: 3000
', '---
components:
- attributes:
    gl/inject-editor: true
  container:
    dedicatedPod: false
    image: registry.gitlab.com/gitlab-org/workspaces/gitlab-workspaces-docs/ubuntu:22.04
    mountSources: true
    env:
    - name: GL_TOOLS_DIR
      value: "/projects/.gl-tools"
    - name: GL_EDITOR_LOG_LEVEL
      value: info
    - name: GL_EDITOR_PORT
      value: ''60001''
    - name: GL_SSH_PORT
      value: ''60022''
    - name: GL_EDITOR_ENABLE_MARKETPLACE
      value: ''false''
    endpoints:
    - name: editor-server
      targetPort: 60001
      exposure: public
      secure: true
      protocol: https
    - name: ssh-server
      targetPort: 60022
      exposure: internal
      secure: true
    command:
    - "/bin/sh"
    - "-c"
    args:
    - ''tail -f /dev/null

      ''
    volumeMounts:
    - name: gl-workspace-data
      path: "/projects"
  name: tooling-container
- name: gl-tools-injector
  container:
    image: registry.gitlab.com/gitlab-org/workspaces/gitlab-workspaces-tools:15.0.0
    env:
    - name: GL_TOOLS_DIR
      value: "/projects/.gl-tools"
    memoryLimit: 512Mi
    memoryRequest: 256Mi
    cpuLimit: 500m
    cpuRequest: 100m
    volumeMounts:
    - name: gl-workspace-data
      path: "/projects"
- name: gl-project-cloner
  container:
    image: alpine/git:2.45.2
    args:
    - |
      echo "$(date -Iseconds): ----------------------------------------"
      echo "$(date -Iseconds): Cloning project if necessary..."

      if [ -f "/projects/.gl_project_cloning_successful" ]
      then
        echo "$(date -Iseconds): Project cloning was already successful"
        exit 0
      fi

      if [ -d "/projects/gitlab-shell" ]
      then
        echo "$(date -Iseconds): Removing unsuccessfully cloned project directory"
        rm -rf "/projects/gitlab-shell"
      fi

      echo "$(date -Iseconds): Cloning project"
      git clone --branch "main" "http://gdk.test:3000/gitlab-org/gitlab-shell.git" "/projects/gitlab-shell"
      exit_code=$?

      if [ "${exit_code}" -eq 0 ]
      then
        echo "$(date -Iseconds): Project cloning successful"
        touch "/projects/.gl_project_cloning_successful"
        echo "$(date -Iseconds): Updated file to indicate successful project cloning"
      else
        echo "$(date -Iseconds): Project cloning failed with exit code: ${exit_code}" >&2
      fi

      echo "$(date -Iseconds): Finished cloning project if necessary."
      exit "${exit_code}"
    command:
    - "/bin/sh"
    - "-c"
    memoryLimit: 1000Mi
    memoryRequest: 500Mi
    cpuLimit: 500m
    cpuRequest: 100m
    volumeMounts:
    - name: gl-workspace-data
      path: "/projects"
- name: gl-workspace-data
  volume:
    size: 50Gi
metadata: {}
schemaVersion: 2.2.0
commands:
- id: gl-tools-injector-command
  apply:
    component: gl-tools-injector
- id: gl-start-sshd-command
  exec:
    commandLine: |
      #!/bin/sh
      echo "$(date -Iseconds): ----------------------------------------"
      echo "$(date -Iseconds): Starting sshd if it is found..."
      sshd_path=$(which sshd)
      if [ -x "${sshd_path}" ]; then
        echo "$(date -Iseconds): Starting ${sshd_path} on port ${GL_SSH_PORT} with output written to ${GL_WORKSPACE_LOGS_DIR}/start-sshd.log"
        "${sshd_path}" -D -p "${GL_SSH_PORT}" >> "${GL_WORKSPACE_LOGS_DIR}/start-sshd.log" 2>&1 &
      else
        echo "$(date -Iseconds): ''sshd'' not found in path. Not starting SSH server." >&2
      fi
      echo "$(date -Iseconds): Finished starting sshd if it is found."
    component: tooling-container
- id: gl-init-tools-command
  exec:
    commandLine: |
      #!/bin/sh
      echo "$(date -Iseconds): ----------------------------------------"
      echo "$(date -Iseconds): Running ${GL_TOOLS_DIR}/init_tools.sh with output written to ${GL_WORKSPACE_LOGS_DIR}/init-tools.log..."
      "${GL_TOOLS_DIR}/init_tools.sh" >> "${GL_WORKSPACE_LOGS_DIR}/init-tools.log" 2>&1 &
      echo "$(date -Iseconds): Finished running ${GL_TOOLS_DIR}/init_tools.sh."
    component: tooling-container
- id: gl-sleep-until-container-is-running-command
  exec:
    commandLine: |
      #!/bin/sh
      echo "$(date -Iseconds): ----------------------------------------"
      echo "$(date -Iseconds): Sleeping until workspace is running..."
      time_to_sleep=5
      status_file="/.workspace-data/variables/file/gl_workspace_reconciled_actual_state.txt"
      while [ "$(cat ${status_file})" != "Running" ]; do
        echo "$(date -Iseconds): Workspace state is ''$(cat ${status_file})'' from status file ''${status_file}''. Blocking remaining postStart events execution for ${time_to_sleep} seconds until state is ''Running''..."
        sleep ${time_to_sleep}
      done
      echo "$(date -Iseconds): Workspace state is now ''Running'', continuing postStart hook execution."
      echo "$(date -Iseconds): Finished sleeping until workspace is running."
    component: tooling-container
- id: gl-project-cloner-command
  apply:
    component: gl-project-cloner
events:
  preStart:
  - gl-tools-injector-command
  - gl-project-cloner-command
  postStart:
  - gl-start-sshd-command
  - gl-init-tools-command
  - gl-sleep-until-container-is-running-command
variables: {}
', 'workspace-url', 'v1', 60, 60, 3, 'devfile-ref', '2025-09-03 13:08:01.069925') RETURNING "id"

Postgres.ai link: https://console.postgres.ai/gitlab/joe-instances/255

MR acceptance checklist

Evaluate this MR against the MR acceptance checklist. It helps you analyze changes to reduce risks in quality, performance, reliability, security, and maintainability.