Skip to content

Releases: finder39/dockge

v2.1.5

Choose a tag to compare

@finder39 finder39 released this 23 Mar 04:21
  • fix: increase agent update timeout from 30s to 5 minutes — large image pulls (e.g. n8n on Unraid) no longer cause 500 errors

Pair with ha-dockge v1.7.2 for the HA integration timeout fix.

v2.2.1b (SSE push notifications)

Pre-release

Choose a tag to compare

@finder39 finder39 released this 08 Mar 22:58

Bug fixes and improvements

  • Fix orphaned SSE operation_started events when stack not found
  • Stop heartbeat timer on server shutdown
  • Extract SSE broadcast helpers for maintainability
  • Add SSE events for down, delete, and auto-update toggle operations
  • Reset heartbeat watchdog on all events (not just heartbeat)

v2.2.0b (SSE push notifications)

Pre-release

Choose a tag to compare

@finder39 finder39 released this 08 Mar 22:34

SSE Push Notifications

Adds real-time Server-Sent Events (SSE) endpoint for the HA integration.

  • GET /api/events — persistent SSE connection for push notifications
  • Events: operation_started, operation_completed, image_check_completed, scheduler_run_completed
  • 30s heartbeat keeps connections alive through proxies
  • All trigger sources emit events: REST API, scheduler, Dockge UI
  • E2E tests for SSE auth, delivery, and failure handling

Pair with ha-dockge v1.8.0 for real-time push updates.

v2.1.4

Choose a tag to compare

@finder39 finder39 released this 09 Mar 03:24
  • fix: add self-stack sidecar to all code paths (deploy, start, REST API update, update-all, scheduler trigger)
  • Previously only the socket handler and scheduler had self-stack detection; the REST API paths would kill Dockge mid-operation when updating its own stack

v2.1.3

Choose a tag to compare

@finder39 finder39 released this 08 Mar 04:47
  • feat: add self-deploy sidecar for deploy/start on Dockge's own stack — deploy and start now safely handle stacks containing Dockge itself using the same sidecar pattern as self-update

v2.1.2

Choose a tag to compare

@finder39 finder39 released this 08 Mar 04:02
  • Fix: Update All and auto-update scheduler now skip stopped/exited stacks instead of attempting to pull and erroring out
  • Adds e2e test coverage for the fix

v2.1.1

Choose a tag to compare

@finder39 finder39 released this 07 Mar 23:10
  • Version display: Show Dockge version for primary and all agents in API (/api/agents), dashboard UI, and HA integration
  • Fix image update detection: Locally-built images (e.g. from quick-deploy) now correctly detect when a newer registry image is available
  • HA integration v1.6.1: New version sensor per agent, removed auto-cleanup of stale devices (users can delete manually)

Paired with ha-dockge-integration v1.6.1.

v2.1.0

Choose a tag to compare

@finder39 finder39 released this 07 Mar 06:01

Changes

Features

  • Persistent ServerAgentManager — the server now maintains its own socket.io connections to all remote agents on startup, independent of browser sessions. The REST API and auto-update scheduler no longer require a logged-in UI session to communicate with agents.

Bug Fixes

  • Multi-agent API reliability — previously, if no one had the Dockge UI open, the REST API could only return stacks from the primary. Now all agent stacks are always available.
  • Auto-update scheduler — scheduled updates for remote agent stacks no longer fail when no browser session is active.

v2.0.2

Choose a tag to compare

@finder39 finder39 released this 05 Mar 01:15

Full Changelog: v2.0.1...v2.0.2

v2.0.1

Choose a tag to compare

@finder39 finder39 released this 04 Mar 21:21

What's New

This is the first release of the Dockge Managed fork. All features below are vibecoded with Claude Code.

Automated Update Management

  • Per-stack auto-update toggle
  • Cron-based scheduler (default: 3 AM daily) with over-run protection
  • Image update detection via skopeo (configurable check intervals)
  • Force check for updates per stack from UI or API
  • Full update history with output logs, filters, and pagination

REST API

  • API key authentication (Settings UI or DOCKGE_API_KEY env var)
  • Stack management: list, status, update, check-updates, auto-update toggle
  • Bulk operations: update-all, system prune
  • Scheduler control: get/set settings, manual trigger
  • Multi-agent support via ?endpoint= parameter

Self-Update Support

  • Stacks containing Dockge itself can be updated (pull + recreate)
  • Destructive operations blocked on self-stack, restart shows warning

Host Path Resolution

  • Auto-detects host stacks directory from container bind mounts at startup
  • No manual DOCKGE_STACKS_DIR_HOST env var needed

UI Enhancements

  • Dashboard: updates available count, next check/auto-update times per agent
  • Compose page: "Image updates available" badge, check updates button
  • Stack list: clock icon for auto-update enabled stacks
  • Update history page at /update-history
  • API key management in Settings

Bug Fixes

  • Fixed frontend/backend version mismatch (nightly Dockerfile stage was default build target)
  • Fixed terminal output race condition for remote agent updates
  • Fixed relative volume path resolution when running in containers

Home Assistant Integration

A companion Home Assistant custom component is available for managing stacks from HA.


Docker image: ghcr.io/finder39/dockge:release