v0.12.1

What's Changed

• Fix interfaces in 0.12 by @lusergit in #1252
• ci: release based on ref name by @lusergit in #1253
• chore: release release-0.12 by @github-actions[bot] in #1254

Full Changelog: v0.12.0...v0.12.1

v0.12.0

0.12.0 (2026-02-18)

⚠ BREAKING CHANGES

• backend: remove authentication bypass configuration

Features

• Add Pause and Resume functionality for campaigns (#1204) (35a9aab), closes #277
• add GraphQL subscriptions for containers (#1224) (6c36e6d)
• add GraphQL subscriptions for OTA operations (#1220) (db679ce)
• Add subscriptions for HT and SMPN (#1225) (7d46654)
• backend: add subscriptions to campaigns domain (2bda866)
• backend: allow deletion of BaseImage used in completed campaign (46f7aa3), closes #598

Miscellaneous Chores

• backend: remove authentication bypass configuration (fb5585c)

v0.11.0

0.11.0 (2026-02-03)

Features

• add calculation to expose Base Image name (d3affa5)
• add device group subscription (#1198) (adb91f4)
• Additional information on deployment events (ba27473)
• backend: base_images subscriptions (bf28b85)
• backend: manual OTA with base image from collection (1af4b8d)
• backend: run tenant reconciliation at creation (f806909)
• GraphQL subscription for device updates in table (#1118) (be6a629)
• implement GraphQL subscriptions for device events (32badc9)
• OTA updates can be canceled (1daed60), closes #266
• show partNumber and serialNumber in Device page if available (#1123) (c2261bb), closes #226

Bug Fixes

• backend: log errors instead of crashing (899f431)
• device reconciliation not working (ba1b025)
• Handle reconciler task replies (bec0758)
• include tenant in socket options (d3c8921)
• unify device creation and update subscriptions (c09e7de)

v0.10.0

This release focuses on adding comprehensive support to manage containerized applications through APIs and the UI. A new application management system allows users to define, deploy, and control custom applications onto managed devices, providing a set of campaigns to make these changes in bulk across device channels.

Several key highlights of this release include:

• Full Container Management System: A complete set of features to define Applications, Releases, Containers, Networks, and Volumes. This is exposed via new UI pages and a full API interface.
• Deployment and Lifecycle Campaigns: Introduction of bulk deployment campaigns (Upgrade, Start, Stop, Delete) that allow users to manage the entire lifecycle of applications on a channel of devices.
• Robust Stability and Device Integration: Significant improvements in deployment reliability through reconciliation logic, manual retry mechanisms, and enhanced integration with Astarte (handling trigger delivery policies and instant device availability for Astarte >= 1.3).
• API and UI Polish: Enhancements to existing features like OTA operation tracking and a major architectural change introducing server-side pagination across many GraphQL APIs to improve performance.

This release includes breaking changes in the GraphQL API due to the migration to Relay/keyset pagination for unbounded lists. Please review the updated GraphQL schema for any adjustments needed in your implementations.

Here is a list of changes since v0.9 (or the previous major version):

Added

• Application Management Feature: Implemented a full application management system, enabling users to view, create, and navigate through Applications, Releases, Containers, Networks, and Volumes.
  • Added dedicated pages for Applications, Releases, Volumes, and Networks (including creation pages).
  • Added an Applications tab to the Device page.
• Deployment Campaigns: Added support for bulk deployment operations:
  • Support for stop campaigns.
  • Support for start campaigns.
  • Support for upgrade campaigns (update).
  • Support for delete campaigns.
• Deployment Reliability:
  • Reconciliation logic: Container information gets reconciled with Astarte if messages get lost.
  • Retry mechanism: Users can manually retry sending necessary deployment messages to a device.
  • Added logic to resume app campaigns when Edgehog starts up.
• Deployment Visibility: Implemented deployments overview with a Deployments page (lists all deployments and enables filtering).
• Infrastructure:
  • The backend now allows defining different images with the same reference but different image credentials.
  • Added support for using Azure Storage as the persistence layer for asset uploads.
  • Added support for trigger delivery policies in the tenant reconciler for Astarte realms (v1.1.1+).
  • Managed OTA operations expose the update target that created them in graphql.
  • Expose the associated UpdateCampaign (if any) from an OTA Operation on the Software Updates tab.
  • Ecto SSL configuration is exposed through DATABASE_* environment variables.

Changed

• BREAKING: GraphQL API that returns unbounded lists now uses Relay/keyset pagination. Edgehog's dashboard now relies on server-side pagination, using tables with infinite scrolling.
• Astarte Integration:
  • Astarte 1.3.0-rc.0 is now the first version allowed for device registration and deletion triggers.
  • With Astarte versions >= 1.3, devices are available in Edgehog upon registration, without waiting for their first connection.

v0.10.0-alpha.9

Added

• Initial draft of the documentation for the container management system.

Changed

• Astarte 1.3.0-rc.0 is now the first version allowed for device registration and deletion triggers.

Fixed

• 422 answer when trying to delete a tenant. The API now correctly handles the tenant_id.

v0.10.0-alpha.8

Added

• Support for stop campaigns: a deployment campaign can send a stop operation on a target release on all devices in a channel.
• Support for start campaigns: a deployment campaign can send a start operation on a target release on all devices in a channel.
• Support for upgrade campaigns: a deployment campaign can update a container to a newer version on all devices in a channel.
• Support for delete campaigns: a deployment campaign can send a delete message on a target release on all devices in a channel.
• Reconciliation logic: container information gets reconciled with astarte if some messages get lost. The polling system acts on time windows to avoid traffic spikes in astarte.
• Retry mechanism: users can manually retry to send necessary messages to a device to deploy a container.

Changed

• The deployment state does no longer incorporate events, instead events are available in the events field of deployments. Users can access them in a log fashion.
• Removed last_message field in deployment.
• With Astarte versions >= 1.3, devices are available in Edgehog upon registration, without waiting for their first connection.

Fixed

• Type error on network and volume options prevented users from using advanced driver options both on volumes and networks.
• Update messages were sent each time a deployment with some ready actions was stopped. This prevented updated releases from actually being stopped.

v0.10.0-alpha.7

Changed

• Added a default image for devices that don’t have an associated system model image, replacing the plain grey placeholder
• Containers now allow to specify binds.
• Frontend validation for port bindings support the full docker specification: docker documentation.
• Adds garbage collection utilities for container resources

v0.10.0-alpha.6

Added

• The backend now allows to define different images with the same reference but different image credentials associated.

Changed

• CPU Period and CPU Quota get validated on the fronted.
• Updated the postgres version in dev, test still run on all versions.
• Added validation for unique volume targets
• Deployment readiness is now a single field
• Deployments now report in the UI the state of each single container. Showing whether the container is running, exited or encountered some error.
• Key-value structure for container environment variables is enforced on backend level.

Fixed

• Reconciler wrongful map comparison.
• Properly update relay connections to avoid out-of sync errors.

v0.10.0-alpha.5

Changed

• When no device group exist, on the Channel List page prompt the user to create at least one group before creating a Channel (#972).
• Improved how Network and Volumes options can be displayed and edited in the Applications section (#963).

Fixed

• Remove application displayed in the application list after deleting it (#967).
• On the device page, propose only options compatible with the device when users want to upgrade an app's release (#970).
• Use uniform and readable labels for displaying Restart Policy values of containers (#971).
• Add logic to resume app campaigns when Edgehog starts up (#969).

v0.10.0-alpha.4

Fixed

• Correctly use Docker's default false value for container's readonly root filesystem option.