Implemented transparent hash version upgrade mechanism that automatically detects and upgrades outdated pipeline version hashes during comparison operations without persisting changes.
Implemented VersionedHash class to provide structured, type-safe access to version hash strings throughout the platform, improving code maintainability and enabling future hash algorithm upgrades.
Removed the pipeline_version_created audit event in favor of the existing pipeline_updated event, which already tracks pipeline version creation.
Added tristate checkboxes to category rows in the custom roles permissions table, allowing users to quickly select or deselect all permissions within a category.
Consolidated Members, Teams, and Collaborators pages into the Access Control page as tabs with pagination support for Teams list and empty-state components.
Added admin endpoint GET /api/admin/orgs/{orgId}/roles to list organization roles.
Added endpoint GET /orgs/{orgId}/users/{userId}/roles to list user roles in organization.
Added maxCustomRolesPerOrg quota definition and enforcement, limiting the number of custom roles per organization.
Fixed Open in new tab functionality for workflow reports by implementing a separate route with report path passed as query parameter.
Fixed step status indication mechanism in launch forms to properly track all controls registered to a section, including those registered after initial creation.
Fixed Data Explorer path decoding to be null-safe, preventing test failures.
Fixed an issue causing custom pipeline info fetch error messages to be incorrectly removed when selecting the same value again in pipeline, revision, or main script fields.
Fixed an issue causing the Launchpad to display empty optimization buttons when optimization status is not set.
Prevent conda Studios from auto-starting when the organization exceeds Studio session limits.
Temporarily disabled Data Explorer integration test that relies on Google Cloud public bucket after access to the bucket changed.
Added Amazon EC2 p6-b300 instance family to the NVMe storage list, enabling support for this new GPU-optimized instance type.
Added the ability to disable and enable compute environments through the UI. This allows administrators or workspace owners to temporarily or permanently prevent new workflow executions or new Studio sessions on disabled compute environments.
Introduced error messaging when attempting to launch workflows on disabled compute environments. The error message now explicitly states Specified compute environment disabled instead of the generic Not available message.
Added Amazon EC2 X8aedz memory-optimized instance family (powered by AMD EPYC processors) to the NVMe storage list, supporting up to 8 TB of local NVMe SSD storage.
Updated the default Azure Batch image and improved error propagation for image-related failures.
Added tracking for Spot instance interruptions (numSpotInterruptions) at the task level, aggregated to the workflow level. This metric tracks how many times Spot instance preemption interrupted and restarted tasks.
Fixed an issue that allowed basic organizations to run more than one Studio simultaneously. Basic organizations now default to a maximum of 1 running Studio when not setting an explicit quota, while other organization types use the global limit (default: 100).
Fixed intermittent workflow launch failures caused by cache returning expired AWS STS credentials. Added expiration checking, configurable refresh margin, and cache invalidation for expired entries.
Reverted Azure Batch to use Ubuntu 22.04 (from 24.04).
Fixed errors when listing compute environments with the resources attribute for legacy Seqera Compute Batch configurations that don't implement IAwsCloudConfig. Added proper handling and fallback logic.
Fixed an issue with parameters not clearing correctly when switching between pipelines with and without parameters in the launch form.
Fixed date formatting to use calendar year (yyyy) instead of ISO week-based year (YYYY), which caused dates at year boundaries to display the wrong year.
Fixed the merging of entity and compute environment labels for single VM compute environments to ensure labels are properly combined and applied.
Improved error messages for insufficient permissions. API 403 Forbidden responses now include details about which specific permissions are missing, making it easier to troubleshoot authorization issues and grant appropriate access.
New AWS Batch compute environments can no longer be created with the deprecated EBS autoscaling feature enabled; use Fusion as the recommended alternative instead. Existing compute environments with EBS autoscaling will continue to work without changes.
Fixed an issue with pipeline optimization being unavailable for shared pipelines in personal workspaces.
Fixed an issue with pipeline optimization permission controls. Users with Launch permissions can now view optimization status and details in read-only mode, while users with Maintain or higher permissions can modify optimization settings.
Added the ability to configure a custom container repository path for Wave-built Studio images in workspace settings, providing greater control over where container images are stored and distributed.
Fixed an issue with interpolated values in dynamic resource labels (e.g., label-name=${workflowId}). Dynamic resource labels now work correctly when enabling the feature for all workspaces.
Fixed an issue with deleted workflows incorrectly counted toward organization quota limits. Active workflow counts now accurately exclude deleted workflows.
AWS Batch compute environments now use AmazonLinux2023-based AMIs by default, offering improved security (IMDSv2), better performance with gp3 EBS volumes, and a more recent kernel. Existing compute environments continue to function with AmazonLinux2 without changes, though we recommend migrating to newer compute environments before AmazonLinux2 reaches end-of-life in June 2026.
Fixed compute environment variables created without specifying a target environment (head or compute), which prevented the system from applying variables during workflow execution. The system now validates that at least one target environment is specified, preventing invalid configurations and providing clear error messages.
Fixed display issues with long helper text in pipeline schema forms. Helper text now wraps properly within form controls, improving readability for parameter descriptions with technical specifications or long value lists.
Fixed an empty launch form issue when launching shared pipelines from another workspace. The launch form now correctly retrieves and displays all pipeline configuration when launching shared pipelines across workspaces.
Fixed OIDC authentication errors with spaces or special characters that caused application crashes instead of displaying proper error messages. Authentication errors during login now display correctly to users, preventing application failures.
Added support for LakeFS Cloud in Data Explorer by enabling image preview and file upload functionality when using LakeFS Cloud as an S3-compatible storage provider.
Improved job failure handling in AWS Batch compute environments. Improved AWS Batch workflow behavior to automatically terminate unschedulable jobs instead of leaving them pending indefinitely, providing faster feedback when compute resources are insufficient for workflow requirements.
Added the g6e instance family to the list of AWS instances with NVMe storage.
Fixed an issue where plots and charts in MultiQC reports appeared as grey boxes instead of displaying correctly, ensuring all visualizations now render properly. MultiQC reports generated by pipeline runs now display complete quality control metrics and interactive charts without missing content.
Fixed an issue where YAML validation errors did not display immediately when importing Studio configurations from Git repositories, ensuring configuration issues are now visible as soon as the import completes.
Fixed an issue where AWS single VM compute environments could not pull container images from private Amazon ECR repositories, bringing them to feature parity with AWS Batch compute environments.
Fixed an issue with older pipeline launch records, where the system did not set the source type properly, resulting in the system failing to access the pipeline entity on launch.
Improved label inheritance when launching workflows via API to automatically apply predefined labels when omitting the labelIds field, aligning API functionality with the GUI and simplifying workflow relaunches.
Fixed an issue where changing the compute environment during workflow resume incorrectly updated the work directory path, causing failures because the workflow could not access cached results from the original run. The work directory now correctly retains the original value when resuming workflows with different compute environments.
Fixed an issue where the settings button in the usage overview widget incorrectly displayed workspace permissions in organization context, and improved workspace settings visibility to dynamically show or hide based on user permissions.
Fixed an issue where workspace settings buttons displayed incorrectly based on user permissions and context, ensuring the settings button now shows the appropriate organization or workspace settings based on the current view.
Added support for date-only formats in search queries, allowing you to search using simple dates (e.g., 2025-02-02) instead of full ISO-8601 datetime strings.
Fixed an issue where filtering for Azure credentials in Data Explorer only returned one credential type instead of all three available Azure credential types.
Fixed an issue where the Data Explorer link did not appear in Task View when configuring multiple AWS credentials with similar permissions in the workspace.
Fixed an issue where Seqera Compute environments did not apply proper EC2 instance type constraints, ensuring consistent compute environment behavior and preventing instance unavailability issues.
Fixed an issue where users could dismiss critical notifications about compute environment substitutions when relaunching pipelines, ensuring important status messages remain visible throughout the workflow.
