Get Started

Put your virtual machines to work on your CI/CD pipeline

Manage and orchestrate multiple Parallels Desktop hosts and virtual machines. Create, start, stop and delete virtual machines and will also allow you to manage the hosts that are running the virtual machines.

Latest release: 0.9.16
Get Started
Learn More
catalog.svg

Run a Catalog service

Securely share golden images with your team quickly.
hosts.svg

Control Multiple Hosts

Manage multiple hosts seamless in a single pane
github_actions.svg

Github Actions

Work with Github Actions and automate your CI/CD pipeline
About

License: Fair Source Builds Release discord

This is the Parallels Desktop DevOps Service, a service that will allow you to manage and orchestrate multiple Parallels Desktop hosts and virtual machines. It will allow you to create, start, stop and delete virtual machines and will also allow you to manage the hosts that are running the virtual machines.

discord.svg

Join our Discord

Join our Discord Server to ask questions, share your experience, and get help from the community.

Join Discord Server
calendar.svg

Schedule a Demo

If you want to schedule a demo, please schedule a meeting with our Solutions Architect. We will go through the service and answer any questions you may have.

Schedule Date
feedback.svg

Leave Feedback

We strive to enhance our services. We kindly invite your to share your feedback regarding your experience.

Leave Feedback
Licensing

You can use the Parallels DevOps service for free with up to 10 users. If you want to continue using the service beyond 10 users, you will need to have a Parallels Desktop Business or Enterprise Edition license.

What's New

Release 0.9.16

Release 0.9.15

  • Enhance GetSizeByteFromString to support additional units and improve error handling

  • Fixes #343

  • Added the missing property for the path where we want to clone

  • Fixed an issue where operations would fail if there was spaces in the vm name

  • Resolve VM registration and retrieval issues

  • Fix machine name being incorrectly set to empty during registration

  • Fix race condition where cached VMs are not updated immediately after registration

  • Added a retry mech to fetch cache vm if it fails

  • switch to manual VM fetching to ensure data consistency after registration

  • Work around prlctl limitation where VM info is incomplete when ID is not specified in command arguments

  • Fixes # (issue)

  • Feature: Added HasWebsocketEvents field to the Orchestrator Host API,providing clients with a reliable source of truth for real-time event capabilities.

  • Improvement: Optimized database performance by implementing lightweight status updates for WebSocket events, avoiding full record rewrites.

  • Fix: Enhanced connection reliability by tying the “Connected” status strictly to verified pong responses rather than just TCP connection establishment.

  • Fix: Resolved delay in reconnection events by triggering an immediate ping upon connection, ensuring instant status updates.

  • Improvement: Refactored connection monitoring to automatically detect and handle stale WebSocket connections without redundant network traffic.

  • bugfix: Associate user with added VM in processVmAdded function

  • docs: add Events Hub user guides and reference documentation

  • Introduces a new documentation section for the Events Hub WebSocket service in docs/docs/devops/events-hub/.

  • Add overview.md: Covers WebSocket protocol details, authentication (Bearer/API Key), connection limits, and available channels (global, pdfm, orchestrator, health).

  • Add “How-to” guides:

  • how-to-monitor-VM-events.md: Subscribing to VM lifecycle events with JSON payload examples.

  • how-to-monitor-orchestrator-events.md: Monitoring cluster-level events like host health and VM state changes.

  • how-to-send-heartbeat.md: Implementing ping/pong logic to keep connections alive.

  • how-to-send-unsubscribe-requests.md & how-to-know-my-client-id.md: Managing subscriptions and client identity.

  • Visualizations: Integrate Mermaid.js todocs: add Events Hub user guides and reference documentation

  • Closes #332

  • Add event-load test command to spawn a dedicated load test server

  • Implement mock auth and realistic PDFM VM event broadcasting (100 events/sec)

  • Register HealthService to support bidirectional ping/pong health checks

  • Add k6 script (loadtest/hub_load.js for capacity testing:

  • Subscribes to both pdfm and health events

  • Tracks latency, message loss, and throughput

  • Verifies bidirectional communication via periodic pings

  • Closes #333

  • New Event Type: orchestrator - Subscribe via /v1/ws/subscribe?event_types=orchestrator

  • Real-time Host Health Updates: Receive notifications when orchestrator checks host health

  • Real-time VM State Updates: Get instant notifications for VM lifecycle events across all managed hosts

  • closes #321

  • Real-time Updates: Eliminates polling delay for hosts with WebSocket support

  • Reduced Load: Significantly decreases API calls for event monitoring

  • Backward Compatible: Gracefully falls back to polling for legacy hosts

  • Scalable: Efficient connection pooling and resource management

  • Resilient: Automatic reconnection and error recovery mechanisms

  • Part of #321

  • Loading the cfg file properly

  • Corrected log messages

  • HealthService: Implemented a singleton HealthService to handle health events (e.g., ping -> pong) using a decoupled Broadcaster interface.

  • SystemHandler: Implemented a singleton SystemHandler to handle system events (e.g., client-id) using a decoupled Broadcaster interface.

  • Optimized Routing: Refactored handleClientMessage to avoid double-marshalling. It now performs partial parsing of the header and passes the raw []byte payload directly to handlers via RouteMessageCmd.

  • Type Safety: Enforced usage of constants.EventType throughout the system

  • Testing: Added comprehensive unit tests for HealthService, SystemHandler, and Client (Reader/Writer/Handler), significantly improving code coverage.

  • closes #319

  • Now all parallels engine events are posted to event’s emitter

  • User’s can now subscribe to PDfM type events and start listening to real time events through a websocket connection.

  • Now Parallels service will broadcast VM_REMOVED,VM_ADDED and VM_STATE_CHANGED to events emitter

  • closes #320

  • After the new event emitter we had a racing condition during the catalog pull process where after registering we would check immediately the presence of a VM but this would not be there due to the event not being fired on time

  • Added a new function to get VMs that gets it sync

  • Updated dependencies on the cache function to this method

  • Created WebSocket connection handler with support for multiple event type - subscriptions.

  • Introduce a type‑safe EventType to replace string literals and prevent processing errors.

  • Added client read/write functions (clientReader/Clietnwriter) for bidirectional WebSocket - communication.

  • Implemented connection limiting per IP address with debug/release build variants.

  • Added events controller with WebSocket subscription and unsubscription endpoints.

  • Added EventEmitter configuration for ping interval and pong timeout settings.

  • Integrated EventEmitter service initialization during application startup.

  • Achieved 57% overall code coverage for the eventEmitter package.

  • related to #319

  • Implemented EventEmitter for managing WebSocket event broadcasting.

  • Created a Hub for handling client connections and message broadcasting.

  • Added Client struct to represent connected WebSocket clients.

  • Introduced methods for sending messages to specific clients, all clients, or by event type.

  • Implemented graceful shutdown for the EventEmitter service.

  • Added tests for environment isolation, EventEmitter functionality, and Hub operations.

  • Included helper functions for setting up test environments and clients.

  • Ensured proper cleanup of environment variables and client connections during tests.

  • related to #319

  • Implemented a new Event Monitor function in the current DevOps Parallels Desktop service.

  • This will read the output of the Event Monitor from prlctl and update the cached list of VMs depending on the state.

  • Now the whole ParallelsService uses event emitter to get real-time status of Vm’s

  • Now in API/Orchestrator will have the real-time VM status

  • PARALLELS_DESKTOP_REFRESH_INTERVAL config is deprecated, vm status is always real-time

  • Fixes 315

  • Fixes 317

  • Removed alludo branding from the documentation

  • Added the new Clarity cookie consent form to the page

  • docs(catalog): add MinIO storage provider details and update push/pull examples with links

  • Updated GitHub links and enhanced folder structure descriptions in source_code page

  • Added missing configs descriptions for pd files

  • Updated configuration documentation with new settings for catalog and reverse proxy

  • Added the ability to disable catalog api credentials obfuscation using env variables (backward compatibility with older clients)

  • Introduced new endpoints for retrieving system logs and streaming logs via WebSocket.

  • Added endpoint to fetch orchestrator host system logs.

  • Implemented endpoint for retrieving orchestrator host reverse proxy configuration.

  • Updated error response structure across various endpoints to use “error” instead of “message”.

  • Enhanced API documentation to reflect the new endpoints and changes in response formats.

  • Added support for compressing manifest files during the generation process.

  • Introduced new fields in the PushCatalogManifestRequest model to handle compression options.

  • Updated the manifest generation logic to calculate original and packed sizes, and log compression details.

  • Added command processors for handling new PDFile commands: COMPRESS_PACK, FORCE, IS_COMPRESSED, VM_REMOTE_PATH, VM_SIZE, and VM_TYPE.

  • Implemented the runImportVM function to handle the import of VMs with the new parameters.

  • Enhanced validation logic to ensure proper handling of new command arguments.

  • Refactored the logic for extracting API documentation from Go files to improve clarity and maintainability.

  • Added module information retrieval to enhance the handling of import paths in the documentation process.

  • Update documentation links to reflect new directory structure for security and catalog sections

  • the /v1/auth/users API response will include a detailed ApiErrorDiagnosticsResponse object. This diagnostics object provides a full call stack plus categorized errors and warnings to help developers identify the cause and the failing component for any request failures.

  • Fixes # (issue)

  • #245