Changelog: Seqera Cloud

Seqera Compute​

Seqera Compute is now in public preview. Seqera Compute is a fully managed, optimized compute environment which eliminates infrastructure complexity by automating environment setup, resource provisioning, scaling, and optimization.

Platform interface updates​

Pipeline run details​

The pipeline run details page has been refreshed and includes improvements such as:

• Clear, dynamic view of pipeline execution in real time via a progress bar.
• A dedicated Run info tab with a centralized view of run metadata and status details.
• Improved navigation tabs for Tasks, Logs, Metrics, and Containers.

New Seqera Platform sidebar​

The Platform top navigation has been replaced with a sidebar for improved navigation.

AWS Cloud compute environment​

The new AWS Cloud compute environment supports faster startup time, simplified configuration, and fewer AWS dependencies. Using this compute environment, Studios can now be launched on Spot instances. The AWS Cloud compute environment is best suited to run Studios and small to medium-sized pipelines. It offers more predictable compute pricing, given the fixed instance type. It spins up a standalone EC2 instance and executes a Nextflow pipeline or Studio session with a local executor on the EC2 machine. At the end of the execution, the instance is terminated.

API documentation improvements​

The Seqera Platform API documentation has been updated. Featuring overview pages with parameter and request body tables, categorized by Platform entity (compute environments, pipelines, etc.), the documentation now provides easier reference and detailed request examples.

Other new features and improvements​

• Fusion
  • Add Fusion snapshots to AWS Batch.
  • Enable DRAINING status for Spot instance reclamation. When an instance is in DRAINING status, no new task can be scheduled on it. See Configuring Amazon ECS Linux container instances to receive Spot Instance notices for more information.
  • AWS Cloud compute environment defaults to the public Seqera AMI if no AMI ID is specified during compute environment creation.
• Studios
  • Add Spot option for AWS Cloud compute environment.
• Support markdown rendering for Nextflow object field help_text and description properties.

Feature updates and improvements​

• Fusion
  • Add Fusion Snapshots support in AWS Batch compute environments, enabling Fusion to automatically restore jobs interrupted by AWS Spot instance reclamation. See Fusion Snapshots for more information.

Bug fixes​

• Fix container table pagination.
• Make task-level containers tab visible only when containers are available.
• Fix a launch form issue with eager parameter keyword interpolation.

Feature updates and improvements​

• Data Explorer
  • You can now drag and drop multiple folders or files to upload them to your cloud bucket, or you can select Upload, then select multiple files to upload.
• Fusion:
  • The Seqera-optimized AMI is now the default for the single VM compute environment. If no AMI is specified when adding a compute environment, it defaults to the Seqera-optimized AMI.
• Add a more detailed warning for workflow cleanup issues that are related to access restrictions to the AWS Batch compute environment.
• Add support for array type parameters in the Launch form.
• Implement support for multiple extension dataset parameters in the input form.
• Delete pipeline secrets in the user context from the provider.
• Switch logs, secrets-manager, marketplace-meter, and simple-email dependencies to AWS SDK v2.
• Show Nextflow config parameters that are binary expressions in the Launch form.
• Add azure_entra to the provider list of credentials in the OpenAPI spec.
• Switch AWS S3 to SDK v2.
• Add task status badges with icons to the workflow run tasks table.

Bug fixes​

• Modify cache expiration policy to resolve "Connection pool shut down" error with new AWS SDK v2 clients.
• Fix an issue where the workflow details page displays the Edit Labels button even when there are no labels, but resource labels are attached to the workflow.
• Fix an issue where the Quick Launch form did not select the primary compute environment by default.

Feature updates and improvements​

• The default Azure Batch Ubuntu image has been updated with sku 22.04. This is to ensure there are no issues with the existing Azure Batch Ubuntu image (sku 20.04) which will be deprecated after April 23, 2025. See, https://github.com/Azure/Batch/issues/174 for more information about the deprecation.

Feature updates and improvements​

• Studios
  • Private sessions: When adding a new session, the number of concurrent connections can now be defined in General Config > Collaboration.
  • Dynamic Studio templates: Labels have been added to indicate the status of support (e.g., Recommended, Deprecated) for a Studio template version. Users can now also switch the Studio's base template image when starting a session as new.
• Nextflow
  • Update WfManifest class to include the new icon property introduced in Nextflow.
  • Add a new migration file to add the new icon column into all the tables of entities that use the WfManifest class as @Embedded property.
• Audit log update: Pipeline edit events are now logged.
• Added totalProcesses count to workflow progress responses.
• Switch AWS Batch compute environment dependencies to AWS SDK v2.
• New workflow job monitoring collects and publishes Platform metrics.
• Support image upload in the pipeline import and pipeline update forms, to be used as pipeline logos.

Bug fixes​

• Fixed a bug where resource labels were erroneously editable for the maintain role when starting an existing Studio.
• Fixed a potential NullPointerException on code related to Google Batch, where the method is expected to return null.
• Fixed an issue where users were not able to remove the organization logo.

Feature updates and improvements​

• Studios
  • Add Studio session lifespan setting. This feature allows you to set a lifespan for a session, after which the session is stopped automatically.
  • Add Studios resource labels support. Workspace admins and owners can edit resource labels inherited from the compute environment and apply them to a Studio session. The edited resource labels associated with each session don’t affect the default resource labels associated with the compute environment.
• Free tier
  • Reduce the maximum number of organization members to 3.
  • Resource tracking in the Platform UI for improved management of free tier limits.
  • User disable/re-enable is now supported.
• Fusion
  • AWS Batch compute environments default to Amazon Linux 2023 AMI when Fusion v2 is enabled
• Implement collapsible view for JSON workflow parameters tab and add View as YAML option.
• Notify the user when secrets cleanup fails.
• Update the pipeline name regular expression to allow pipeline names containing dots ('.').
• Mark long-running jobs with a warning: This warning will be displayed when a workflow has been in the RUNNING state for more than 3 hours and no single task moved from one state to the next.
• Set ECS_CONTAINER_START_TIMEOUT to 10 minutes in the ECS settings used by AWS Batch CEs in order to prevent the error Task failed to start - DockerTimeoutError: Could not transition to started; timed out after waiting 3m0s.
• Upgrade to Angular 18

Bug fixes​

• Studios
  • Fixed an issue with searches for names containing special SQL wildcard characters (_, %).
  • Fixed an issue to ensure that the Maintain role is not able to change resource labels inherited from the selected compute environment when adding/starting a Studio.

Feature updates and improvements​

• Studios
  • Display the applied date filter on the Dashboard page.
  • Add support for custom templates for input control options such as workspace visibility and session lifespan.
• Fusion
  • Default to Amazon Linux 2023 AMI when Fusion v2 is enabled.
  • Add support for Fusion licensing.
• (Enterprise only) Add a job to save usage metrics to files so usage data can be accessed and shared by users even when they’re not connected to License Manager.
• Implement FaviconService that changes the favicon based on pipeline run or Studio status.
• Increase accessTokenExpirationTime by 10% to always trigger a token refresh when Platform is open in a browser tab that isn't hibernated.
• Allow Nextflow configuration parameters with embedded references to other parameters to be shown verbatim.
• Extend zombie jobs detection on all workflows reaching one of the terminal states (SUCCEEDED, FAILED, CANCELLED).

Bug fixes​

• Fixed an additional reference to the mainScript parameter in the pipeline-info API response.
• Improved security of passthrough redirection on Platform authentication.

Breaking changes​

• Restore errorStrategy and maxRetries default Nextflow config for AWS and GCP launches. Platform will not overwrite existing error strategy coded in the pipeline repository.
• If you want to automatically retry on task failures (e.g., from spot reclamations), you will now need to explicitly set errorStrategy and maxRetries.
• Automatic spot retries in AWS and GCP need to be actively enabled. Alternatively the generic failure is handled at the Nextflow level.

You can revert to the default setting used in Nextflow prior to 24.10 with the following configuration options:

aws.batch.maxSpotAttempts = 5
google.batch.maxSpotAttempts = 5

Set task.maxRetries to a non-zero value to restore Platform to prior behaviour.

How do I retry Spot failures while making them visible?

There is no mechanism in Nextflow to specifically report Spot retries. However you can set process.maxRetries = 5 to retry all failures (including Spot reclamations) a specified number of times before failure.

Feature updates and improvements​

• Data Studios:
  • Added data studio API endpoints to the OpenAPI spec.
• New dynamic page title for easy bookmarking.
• Free tier:
  • New organization types added to allow free tier users to monitor their usage and limits.
• Restore errorStrategy default Nextflow config for AWS and GCP launches. The Nextflow job now correctly terminates after pipeline completion.
• Improved error messaging when pipeline info can't be fetched indicates whether the issue is due to a missing resource, or failed authentication due to expired credentials.
• Azure jobs are now automatically terminated after all tasks are complete.
• Send only added/updated run parameters when launching a pipeline. This includes all defaults and parameters passed during the launch.
• Updated list of EC2 families with NVMe disks available.
• Upgrade to Angular 17
• Bump Nextflow launcher to 24.10.3

Bug fixes​

• Allow unsetting schemaName when editing a pipeline.
• Use preferred_username as fallback email field for OIDC login.
• Drop the last characters in job definition name, if longer than the prescribed limit.
• Compute environment creation form allows pre- and post-run scripts to be longer than the accepted value of 1024 characters.
• Prohibit duplicate Git credentials and tie-break on lastUpdated for existing duplicates. Going forward, preference is given to the credentials updated last in the situation when there are multiple candidates. Additionally a duplicate check is added when creating new credentials, to avoid new duplicate cases.
• Disable reset selection on data change in workflow list component.
• Take into account the alternative mainScript path.

Feature updates and improvements​

• Add AWS Forge support for Amazon Linux 2023.
• Deprecate Google Life Sciences API.
• Refactor reports file streaming over Agent connections.
• Bump Nextflow launcher to 24.10.0.
• Add name validation to custom containers during Data Studio session creation.

Bug fixes​

• Propagate EFS mount path to container volume.
• Include optimization data to launch request, if present.
• Correctly update user through profile page.
• Input form view not working for private GitHub repos during pipeline quicklaunch.
• Improve layout on large displays.
• Add missing permissions check to data studio checkpoint "Start as new" option.
• Workspace secrets not selected by default during pipeline quicklaunch with new launch form.
• Add volumes to data studio job definition.
• Users with "Maintain" role and lower can edit resource labels when launching with the new pipeline launch form.
• Do not register the "User" route if user personal workspace is disabled.
• Apply the old launch form behavior to the new launch form when launching a shared pipeline from a private workspace.
• Improved error messaging for expired Git credentials.

Seqera Cloud authentication updates​

Seqera Platform Cloud has migrated to a new authentication provider, which includes two major changes to the login flow.

Passwords for email login​

• Users signing in via email: The next time you sign in to Seqera Cloud, select the option to create a new account — "Create one" — and add a password for your existing account. All your existing account information will be preserved and all future sign-ins can be completed with your new password.

• Users using Github/Google sign-in: You are not required to define a password for your account and shouldn’t see any changes on your next sign-in.

Account linking​

Users that sign in with multiple accounts associated with the same email address: You will be prompted to optionally merge your accounts into a single account. This will merge all your work across the two different accounts into the same account and give you the option to sign in via either method of authentication.

If you choose not to merge accounts, you will not be prompted again. If you wish to merge accounts later, or would like to merge accounts with different email addresses, contact Seqera Support.

Feature updates and improvements​

• Data Studios now support GPUs.
• The pipeline parameters input form now supports JSON schema draft 2020-12.
• Augment Data Studios images with custom Conda environment.
• Data studio creation now includes YAML validation.
• Data Studios build stats are now included in the Dashboard.
• New data studios now include "Add only" and "Add and start" actions.

Bug fixes​

• Add back to login button in authentication error screen.
• Add custom token cookie login handler for Nextflow requests to receive authentication cookies.
• Data Studios UI fixes.
• Jedis connection pool config which caused potential instabilities.
• Improve Data Studios error messaging.
• Delete Data Studio checkpoint on build failure.
• Update default Data Studios tool versions.
• Failed Data Studios can now be deleted.
• Set launch form to parameters step view by default when a pipeline schema is present, regardless of specification.
• Dashboard persisted filters exclude workspace IDs the user no longer has access to.
• Do not fetch info for INVALID status pipelines, fetch for all other statuses including DISABLED.
• Fixed a problem with Azure Entra credential creation.
• Return an empty list of data studios when receiving an invalid status.
• Fallback to primary compute environment when launching a shared pipeline without an associated compute environment from a private workspace.