Compute environments​

• Added the compute environment ID to the compute environment row card.
• Added the c9gd instance family to the NVMe instance list.
• Added a force query parameter to the credentials POST /validate endpoint for operator override.
• Updated the list of available Azure regions.

Bug fixes​

Compute environments​

• Added a clear error message when Azure Batch pool deletion fails.
• Disabled compute environment form submission while the form is invalid.
• Classified SDK-wrapped network errors as transient in the credentials preflight cron.
• Backed off pre-flight compute environment validation on transient or error probes.
• Accepted cross-account allowBuckets work directories in pre-flight checks.

Data explorer​

• Signed GCS genomic file URLs with V4 for IGV preview.

Access control​

• Skipped non-RSA tokens in the OIDC verifier to reduce log noise.

General​

• Added SSRF host validation using lib-util-net by @pditommaso in #1061
• Added the waveBuildNotification user preference for build email notifications by @robnewman in #1069

Bug fixes​

Container building​

• Fixed conda/micromamba:v2 builds to use the micromamba 2 image by @pditommaso in #1065

Full changelog: https://github.com/seqeralabs/wave/compare/v1.34.0...v1.35.0

Language features​

• Clarified storeDir caveats with env, eval, and stdout outputs by @bentsherman in #7235

Azure​

• Added docs tip for relocating Docker storage on Azure Batch nodes by @adamrtalbot in #6937

Fusion​

• Added a clear error when Fusion is enabled without an access token by @jorgee in #7263

Plugins​

• Consolidated the nf-tower plugin into nf-seqera by @pditommaso in #7252

General​

• Added Fibonacci and Pareto to the run name generator by @bentsherman in #7255
• Normalized version callouts in the docs by @bentsherman in #7256
• Propagated the Seqera Intelligent Compute scheduler run identifier to Platform by @jorgee in #7228
• Removed uncommitted specs by @bentsherman in #7267

Dependencies​

• Bumped nf-seqera@1.0.0

Bug fixes​

Language features​

• Fixed docs example for record operations by @bentsherman in #7250
• Fixed flatMap tuple error for dataflow value by @bentsherman in #7254

General​

• Fixed failing unit tests with clean build by @bentsherman in #7264
• Lowered trace-access.nf memory to fit CI runner capacity by @pditommaso in #7266

Full changelog: https://github.com/nextflow-io/nextflow/releases/tag/v26.06.0-edge

Compute environments​

• Added VPC and multiple subnet selection to AWS Cloud compute environments.
• Added the number of instances provisioned per run to the run page, sourced from the Scheduler run summary.
• Added the g7 instance family to the NVMe list.

Pipelines​

• Updated Nextflow to v26.04.4.

Access control​

• Split the data_link and data_link_object permissions into separate scopes. Workspace members with View, Connect, or Launch roles can now download files and generate download URLs from data-links. Upload and delete file operations remain restricted to Maintain and above.
• Added pre-filling of first and last name from GitHub signup.

Monitoring and observability​

• Added audit entries for pre-flight cron status transitions.

General​

• Added a default workspace preference to the user profile.
• Listed avatar and username rows in the avatar-group popover.

Bug fixes​

Compute environments​

• Fixed AWS STS to allow instance credentials without an explicit role mode.
• Fixed the AWS Cloud head job to honor vpcId and subnet network selection.
• Downgraded ecs:DescribeContainerInstances permission errors to WARN level.

Pipelines​

• Fixed lineage incorrectly showing as not enabled for submitted runs when it was enabled.

Access control​

• Fixed detection of a personal access token as a duplicate when creating a GitHub App with the same credentials.
• Fixed domain verification to be enforced during SSO enablement and ticket refresh.

Pipelines​

• Added pipeline schema, commit ID, head job CPUs, and head job memory to the workflow launch form summary, plus pipeline version and lineage fields when those features are active.

Bug fixes​

Compute environments​

• Fixed AWS credential validation failing for credentials using assumeRoleArn when TOWER_ALLOW_INSTANCE_CREDENTIALS=true without an explicit role mode.

Pipelines​

• Fixed duplicate credential detection to block GitHub App creation when a personal access token credential already exists for the same repository base URL.

Upgrade notes​

No breaking changes. Standard upgrade procedure applies.

Language features​

• Removed the deprecated echo process directive by @pditommaso in #7206
• Added support for record type input with dot-notation CLI params by @jorgee in #7170

Azure​

• Improved Azure Batch VM image diagnostics and fixed stale 24.04 docs by @pditommaso in #7200

Plugins​

• Added workflow DAG reporting to the nf-tower plugin by @ewels in #7218
• Allowed the plugin version to be specified when running a plugin command by @JosuaCarl in #7113
• Added subcommand listing to nextflow plugins -help by @ewels in #7197

General​

• Added -help as an alias for -h in top-level CLI options by @ewels in #7238
• Added GPU metrics to the workflow-level execution summary by @cristianrcv in #7229
• Added hidden plural/singular aliases for noun subcommands by @ewels in #7239
• Added Nextflow module info to the lineage TaskRun entry by @jorgee in #7160
• Added seqera:// data-links support to the nf-tower filesystem by @jorgee in #7070
• Added a shellEnabled option and per-task prediction model to the Seqera executor by @pditommaso in #7204
• Added the nextflow logfile command by @ewels in #7177
• Added a data lineage architecture decision record (ADR) by @bentsherman in #7187
• Changed stdout outputs to stderr to avoid breaking piped output by @ewels in #7121
• Clarified the distinction between modules and script inclusion in the docs by @bentsherman in #7217
• Migrated the docs to the new build by @christopher-hakkaart in #7119
• Improved Tower telemetry error handling on transient gateway failures by @pditommaso in #7190
• Mirrored test containers to public.cr.seqera.io to harden CI against quay.io outages by @pditommaso in #7195
• Promoted the latest stable channel by version comparison instead of branch name by @pditommaso in b5ac62e0c
• Refactored changelog.txt to CHANGELOG.md by @ewels in #7179
• Added support for the accelerator directive in the local executor by @bentsherman in #5850
• Used a stop flag instead of a thread interrupt to shut down the logs checkpoint by @pditommaso in #7193

Bug fixes​

Language features​

• Fixed "No such variable" in typed output block with job arrays by @jorgee in #7221
• Fixed cast and instanceof with record types by @bentsherman in #7185
• Fixed staging of Path fields of included record types by @robsyme in #7226

Plugins​

• Preserved the order of plugin declarations in config by @jagedn in #7227

General​

• Fixed workflow completion handler being called twice by @bentsherman in #7191
• Fixed lineage records for stub-run and eval outputs by @jorgee in #7211
• Fixed netty CVE-2026-44249 via an AWS SDK bump and azure-core-http-netty pin by @jorgee in #7223
• Fixed nxf_date for uutils coreutils (Ubuntu 26.04+) by @robsyme in #7118
• Fixed nxf_date to sample seconds and nanoseconds atomically by @pditommaso in #7207

Dependencies​

• Bumped nf-amazon@3.10.0
• Bumped nf-azure@1.23.0
• Bumped nf-codecommit@0.6.0
• Bumped nf-seqera@0.22.0
• Bumped nf-tower@1.29.0

Full changelog: https://github.com/nextflow-io/nextflow/releases/tag/v26.05.0-edge

Compute environments​

• Redesigned the compute environment list as a row-card table.
• Added attribute chips to the compute environment list.
• Added an on-demand credentials validation button.

Bug fixes​

Pipelines​

• Surfaced launch-form errors on disabled fields so pipeline-info errors (bad URL, missing main script, unknown revision) no longer disappear silently.
• Preserved optimization targets verbatim on workflow launch.

Plugins​

• Bumped the default plugin create template to v0.4.0 by @pditommaso in #7234

General​

• Added Nextflow module info to lineage TaskRun entry by @jorgee in #7160

Bug fixes​

Language features​

• Fixed "No such variable" in typed output block with job arrays by @jorgee in #7221
• Fixed cast and instanceof with record types by @bentsherman in #7185
• Fixed staging of Path fields of included record types by @robsyme in #7226

Plugins​

• Preserved order of plugin declarations in config by @jagedn in #7227

General​

• Fixed workflow completion handler being called twice by @bentsherman in #7191
• Fixed nxf_date for uutils coreutils (Ubuntu 26.04+) by @robsyme in #7118
• Fixed nxf_date to sample seconds and nanoseconds atomically by @pditommaso in #7207

Dependencies​

• Bumped nf-amazon@3.9.2
• Bumped nf-azure@1.22.6
• Bumped nf-codecommit@0.5.3
• Bumped nf-seqera@0.21.0
• Bumped nf-tower@1.28.1

Full changelog: https://github.com/nextflow-io/nextflow/releases/tag/v26.04.4

Compute environments​

• Added AWS Cloud compute environment EBS volume encryption with an optional KMS key.
• Added an INVALID-status error alert on the credentials view page.
• Added display of the AWS jump-role trust policy.
• Added the m9gd instance family to the NVMe list.
• Relaxed the credentials validation cron interval to 12 hours.

Pipelines​

• Added a launch output directory option to the launch form.
• Added a shareable deep link to the task details view.
• Showed the metrics tab for canceled workflow runs.

Bug fixes​

Compute environments​

• Downgraded a SLURM/HPC executor error to a warning so transient submission issues no longer surface as failures.
• Treated AWS SDK network errors as transient so they no longer fail compute environment operations on the first retry-able blip.

Monitoring​

• Included gpuMetrics in task JSON serialization so GPU usage is now visible in the task payload.

Pipelines​

• Fixed an unexpected error when the workflow limiter canceled a newly launched workflow.
• Preserved a custom schema when added for the first time in pipeline edit mode.

Container building​

• Made micromamba/v2 the default build template for Conda packages by @pditommaso in #1024

Dependencies​

• Bumped io.netty:netty-bom to 4.2.13.Final by @cristianrcv in #1053

General​

• Made the server listening port configurable via the WAVE_SERVER_PORT environment variable by @bebosudo in #1029
• Made the STDOUT log pattern configurable via WAVE_LOG_STDOUT_PATTERN by @munishchouhan in #1055
• Removed git tag --force from the Wave release CI by @ShahzebMahmood in #1056
• Removed the legacy wave/app ECR push by @ShahzebMahmood in #1057
• Used an OIDC role for Maven publishing by @ShahzebMahmood in #1059
• Added Wave features documentation by @christopher-hakkaart in #1040

Bug fixes​

General​

• Ported the Wave CI release flow with a non-enterprise jibRepo fix by @ShahzebMahmood in #1046

Full changelog: https://github.com/seqeralabs/wave/compare/v1.33.6...v1.34.0

Azure​

• Updated default Azure Batch pool image to Ubuntu 24.04 by @pditommaso in 114d0e4c

General​

• Pinned rnaseq-nf to v2.3 in Azure validation script by @pditommaso in 8f7e5780

Bug fixes​

S3​

• Fixed S3FileSystemProvider.newInputStream() draining full object on close by @jorgee in #7046

General​

• Fixed world-readable seqera-auth.config credential file by @arnaualcazar in #7182
• Fixed LogsCheckpoint thread interrupt race condition by @pditommaso in #6939

Dependencies​

• Bumped nf-amazon@3.4.5 by @pditommaso in b9074b2d
• Bumped nf-tower@1.17.7 by @pditommaso in da09f092
• Bumped nf-azure@1.20.3 by @pditommaso in 1e000afe

Full changelog: https://github.com/nextflow-io/nextflow/releases/tag/v25.10.6

Access control​

• Enabled GitHub App credentials for all workspaces.

Compute environments​

• Added security group name search to AWS drop-downs.

Pipelines​

• Surfaced launch-time settings, including profiles and syntax parser, in the run config tab.

Bug fixes​

Data explorer​

• Fixed IGV index file handling by presigning indexURLs independently.

Pipelines​

• Fixed the Nextflow syntax parser v2 toggle to seed from the saved launch config.
• Included syntaxParser in the workflow launch API response.
• Cleared stale main script launch errors.
• Fixed the launch form to skip the lineage request in user space.

Language features​

• Improved the error message when enabling types with syntax parser v1 by @jorgee in #7173

Plugins​

• Made plugin metadata prefetch resilient to registry rate limiting by @pditommaso in #7181

Fusion​

• Added support for fetching local files in FUSION_CONTAINER_CONFIG_URL by @alberto-miranda in #7142

Bug fixes​

S3​

• Fixed bucket cleanup in S3 test by @bentsherman in #7172

General​

• Fixed docs CI on Python 3.14 by upgrading Sphinx by @jorgee in #7175

Dependencies​

• Bumped nf-azure@1.22.5
• Bumped nf-google@1.27.3
• Bumped nf-k8s@1.5.4
• Bumped nf-seqera@0.20.1
• Bumped nf-tower@1.28.0
• Bumped nf-wave@1.21.0

Full changelog: https://github.com/nextflow-io/nextflow/releases/tag/v26.04.3

Pipelines​

• Exposed tower.launch.config.maxSize to frontend validators so launch configuration size limits stay aligned with the backend.

Bug fixes​

Pipelines​

• Added frontend validation of paramsText against its JSON-serialized size to align launch parameter limits with backend submission constraints.
• Added missing metadata, pipeline versioning, and data lineage fields to the workflow launch form summary.
• Fixed the workflow launch payload to always send lineage.enabled so the data lineage toggle setting persists correctly.

General​

• Fixed the workspace switcher to keep the search input focused during arrow-key navigation.
• Fixed background worker shutdown by draining workers before JedisPool shutdown to prevent dropped jobs.
• Fixed navigation switcher responsiveness.

Highlights​

• Co-Scientist is now available for Enterprise deployments.
• Pipelines now support a GitHub App manifest flow for credential creation, with higher throughput than personal access tokens (PATs).
• Compute environments support descriptions and can be disabled, and Google Batch and Azure Batch have significant stability and scalability improvements.
• AWS, Azure, and GCP each gain deeper credential and identity management, including Workload Identity Federation (WIF) for Google Cloud Batch and Entra service principal support for Azure Batch.
• Nextflow data lineage is configurable and attached to pipeline runs and Data Explorer objects.
• Existing Studios can be edited, and new workspace settings for custom container registry controls and default collaboration modes provide top-down granular management.
• Datasets now support linked (URL-referenced) remotely-hosted data.
• Data Explorer adds IGV and Molstar 3D viewer for genomic and structural files, respectively.
• Audit logs have a new user interface with on-demand CSV export, configurable maximum records, and richer event context. The Admin panel exposes real-time usage metrics.

Feature updates and improvements​

Co-Scientist​

• Added support for Co-Scientist in Enterprise deployments.

Studios​

• Enabled Studios on all workspaces by default. See the Upgrade guide for more information.
• Added workspace setting for default Studios session collaboration.
• Added multi-stage builds and Wave v1.33.0 support for custom Studios containers.
• Added nameStrategy configuration option to Studios workspace settings.
• Added ability to edit the settings of stopped Studios.
• Container registry improvements, including:
  • Configurable container repository path for custom image builds.
  • Configurable container image naming strategy (default, tag prefix, image suffix, none).
• Added new environment variables. See Data features for the full descriptions and defaults:
  • SSH settings:
    • Allowed workspaces
    • Enable key management
    • Limit SSH connections to authenticated sessions via key fingerprint
    • Custom SSH server domain
    • Custom connection string display port
  • Custom Studio images:
    • Disallowed image repositories
    • Image naming strategy
    • Wave build status checks
  • Default privacy settings (private vs. collaborative by default)
  • Studios displayed per page
  • Feature manifest URL for version compatibility checks
  • Per workspace startup metrics & retention policy

Compute environments​

• Added ability to disable a compute environment.
• Improved compute environment form warning display with individual stacked alerts.
• Added m8gn, m8gb, c8id, m8id, r8id, x8aedz, g7e, p6-b300 instance families to NVMe list.
• Added new environment variable. See Configuration overview for the full descriptions and defaults:
  • TOWER_COMPUTE_ENV_CLEANUP_*

Azure​

• Added separate head and worker pool support for Azure Batch compute environments in both Forge and manual modes.
• Changed default Azure Batch job timeout to 7 days and exposed it as a new configuration item.
• Updated default Azure termination policy in compute environment creation form.
• Added VNet and subnet support for Azure Batch compute environments.
• Added support for separate managed identity client IDs and resource IDs for head and compute jobs in Azure Batch.
• Enabled Entra service principal credentials for Azure Batch Forge pool creation and Fusion v2.

AWS​

• Added AWS credential modes with support for key-based and role-based access.
• Added AWS External ID support for role-based credentials.

GCP​

• Added Workload Identity Federation (WIF) credential support for Google Batch and Google Cloud compute environments.
• Added support for network tagging in Google Batch.
• Added boot disk image selection for Google Batch compute environments.
• Added support for multiple machine types in Google Batch compute environments.
• Added configurable retry behavior for tasks with null exit codes in Google Batch compute environments.

Pipelines​

• Redesigned workflow notification email templates with updated styling.
• Added GitHub App manifest flow for credential creation.
• Improved clipboard UX in workflow details header.
• Redesigned report preview modal header layout and modal.
• Registered Nextflow CLI as a static OIDC client for authorization code with PKCE flow.
• Enriched the POST /trace/create response with platform metadata to reduce downstream API calls from Nextflow.

Datasets​

• Added preview for linked (URL-referenced) datasets.

Data Explorer​

• Added IGV for genomic data visualization.
• Added Molstar 3D viewer for molecular data visualization.
• Added extensible view mode selection for JSON files in Data Explorer (JSON, IGV, and plain text).
• Added auto-detection for IGV-renderable files (.bam, .bed, etc.).
• Updated Data Explorer to display non-native browser files as text when opened in a new tab.
• Added Fusion symlink resolution to the Data Explorer API.
• Increased the maximum data link name length to 512 characters.

Access control​

• Added required description field to custom role creation.
• Exposed roles API endpoints in the OpenAPI specification.
• Added SSO domain-based redirect for the login guard.

Monitoring and observability​

Admin panel

• Added real-time active user count display.
• Added workspace usage metrics.

Audit logs v2: new capabilities

• Added CSV export from the admin audit logs v2 table, with a configurable maximum record limit.
• Switched to token-based pagination for improved performance.
• Added comprehensive logging for SSO lifecycle events.
• Added descriptions and documentation metadata to audit event types, and renamed outdated event types for consistency.

Audit logs v2: richer event context

• Added owner ID and workspace ID metadata to Studios session audit events and to all data link audit events.
• Added target organization, workspace, and user context to audit log v2 interfaces.
• Added a target_name field to the data model and surfaced target resource names across event emission points, the admin table, the API, the UI, and CSV exports.

Audit logs v2: migration and cleanup

• Migrated telemetry usage queries to the v2 table.
• Updated the audit log cleaner to handle both v1 and v2 tables.
• Removed unused instanceId and instanceName columns from the v2 table.

Environment variables

• See the Configuration overview for the full descriptions and defaults:
  • TOWER_AUDIT_LOG_V2_WRITE_MODE
  • TOWER_AUDIT_LOG_V2_CSV_EXPORT_MAX_LOGS
  • TOWER_AUDIT_LOG_V2_PRE_POST_CHANGE_ENABLED
  • TOWER_CRON_AUDIT_LOG_CLEAN_UP_*

General​

• Bumped Micronaut from 4.7.6 to 4.8.3.
• Improved admin workspace list toolbar responsiveness.
• Applied updated status icons across platform components.
• Redesigned page header layout with improved toolbar and breadcrumb integration.
• Added automatic breadcrumb navigation to page headers.
• Updated delete workspace confirmation modal text to clarify the impact of deletion.
• Removed the dynamic resource labels feature toggle (feature is now always active).

Bug fixes​

Studios​

• Added workspace existence check before creating Studios workspace settings.
• Added validation of Git repository configuration files when creating a Studio.
• Fixed broken navigation from Studio details page for private Studios.
• Updated R-IDE to respect compute environment and Studio-specific environment variables.
• Fixed CSP header empty directive.

Compute environments​

• Fixed Google Batch machine type migration to be portable across MySQL and MariaDB.
• Removed hardcoded prediction model configuration from AWS Cloud platform provider.
• Fixed metering event handling to batch events when more than 100 events are received, preventing silent data loss.
• Removed default terminateAsync implementation to enforce explicit provider implementations.
• Propagated AWS Forge disposal failures instead of silently swallowing exceptions.
• Pinned google-cloud-storage to a compatible version to fix NoClassDefFoundError on GCS data link access.
• Returned an actionable error message when an Azure Batch pool is missing during job submission.
• Propagated GCP Forge disposal failures instead of silently ignoring resource deletion errors.
• Enabled cloud cache for Kubernetes compute environments with local PVC paths.

Pipelines​

• Fixed pipeline implicit default version resolution.
• Removed logs from AI debug button URL to prevent URI too large errors.
• Replaced document.write with client-side form submission in GitHub App manifest flow to fix Firefox blank page issue.
• Made workflow job cancellation idempotent to prevent 500 errors on concurrent cancel requests.
• Fixed parallel requests to pipeline info in the launch form.

Datasets​

• Fixed LazyInitializationException in avatar resolution during dataset creation.
• Fixed dataset name field to apply input normalization (spaces converted to underscores).
• Fixed column order preservation in dataset preview for TSV files.

Data Explorer​

• Fixed IGV MIME type detection in Data Explorer.

Monitoring and observability​

• Fixed dashboard drop-down scrolling and character overflow.
• Fixed task logging to use populated taskId instead of empty id.
• Fixed user_sign_in audit events to correctly populate the actor field with the signing user's ID.

General​

• Fixed side navigation width not updating in Safari when toggling the collapsed state.
• Fixed credits banner appearing during page load.
• Fixed oversized icon on the forbidden access page.

Upgrade notes​

Breaking changes​

Audit log versions in 26.1​

Seqera Platform Enterprise 26.1 introduces the audit log v2 schema as a breaking change for direct database consumers and custom ETL jobs. See the Upgrade guide for guidance.

Azure Batch Ubuntu 24.04 and AppArmor​

The ubuntu-hpc image in Azure Batch now defaults to Ubuntu 24.04, whose kernel and AppArmor policy block the unprivileged user namespaces and FUSE mounts Fusion needs. Running Fusion on Ubuntu 24.04+ pools therefore requires an AppArmor profile that permits Fusion mounts to be loaded on every node, plus a matching --security-opt on each task container. This requires installing the AppArmor profile on the pool nodes. See Azure Batch AppArmor.

Forge compute environments

Seqera Platform handles everything automatically. No action required.

Manual compute environments

When Fusion is enabled and the pool's image SKU is 2404 or higher, Seqera Platform automatically appends the options below to the generated Nextflow configuration:

process.containerOptions = '--security-opt apparmor=seqera-fusionfs-container'

Compute environments​

• Added a configurable boot disk size for Azure Batch pools.
• Added r8idn, r8idb, m8idn, and m8idb instance families to the NVMe list.

Monitoring​

• Improved the JSON diff component with small refinements.

Pipelines​

• Added the Lineage ID to the Run Info page.

General​

• Introduced a panel search input and integrated it into the navigation switcher component.
• Improved workspace switch redirects to preserve the active sidebar section.

Bug fixes​

Access control​

• Fixed the audit event type for MANUAL-to-SCIM promotion.
• Kept the infinite-scroll drop-down open when clicking the search input.

Compute environments​

• Improved Entra robustness by adding equals/hashCode for credentials and making the managed identity regex case-insensitive.
• Deduplicated the compute environment entry in the launch drop-down.

Data Explorer​

• Made credentials a required field in the Add data repository form.
• Fixed the Data Explorer custom bucket E2E test to select credentials.

Monitoring​

• Fixed a referential integrity constraint violation when saving an audit log image bundle in H2.
• Replaced the word ID with IP in the audit logs table.
• Coalesced setAvailableCredits calls per organization.
• Coalesced stop events per organization using a Redis lock.
• Added a differential TTL and empty-result guard for the Metronome balance cache.
• Coalesced handleOrganizationCreditsExpiration calls per organization.
• Skipped null balance puts in listGrantBalance to avoid a ConcurrentHashMap NPE.

Pipelines​

• Refined workflow_updated events that contain no diff.
• Prevented duplicate GitHub credentials across PAT and GitHub App types.
• Resolved file paths to the most recent lineage record.

General​

• Fixed the expandable-text overlay to display only when text is truncated.
• Preserved panel search focus inside mat-menu.

Language features​

• Validated module spec against JSON schema by @pditommaso in #7094
• Updated guidelines for strict parser by @bentsherman in #7134

General​

• Honored token field in BitbucketServerRepositoryProvider by @jorgee in #7120
• Picked up force-pushed branches in multi-revision repo strategy by @jorgee in #7103
• Tightened CI security and removed dead code in workflows by @ewels in #7139
• Updated GitHub Actions with security updates by @ewels in #7138

Dependencies​

• Pinned python to 3.14.5 by @seqeralabs-renovate in #7152
• Bumped nf-azure@1.22.4 by @pditommaso in 2bcc6e8c
• Bumped nf-k8s@1.5.3 by @pditommaso in b2ea7492
• Bumped nf-seqera@0.20.0 by @pditommaso in 4d55e81e

Bug fixes​

Language features​

• Fixed module run to allow optional path inputs by @bentsherman in #7163
• Fixed destructured record input in module run by @bentsherman in #7164
• Fixed function calls in typed process outputs by @bentsherman in #7148
• Fixed strict parser to include params refs in task hash by @bentsherman in #7165
• Normalized TaskPath to Path in typed process outputs by @stekaz in #7090

Azure​

• Fixed Azure rnaseq-nf test by @bentsherman in #7143

General​

• Released HTTP sockets when reading HTTP attributes to avoid 503 throttling by @jorgee in f2b0284e

Full changelog: https://github.com/nextflow-io/nextflow/releases/tag/v26.04.2

General​

• Demoted 'Request processing' log to trace in RegistryProxyService by @pditommaso in #1047

Full changelog: https://github.com/seqeralabs/wave/compare/v1.33.5...v1.33.6

General​

• Ported Wave CI to push to internal ECR (SEC-1395) by @ShahzebMahmood in #1039
• Embedded OpenAPI spec and Swagger UI in main Wave container by @pditommaso in #1043
• Used addIfLess for JWT refresh timer scheduling by @pditommaso in #1045

Dependencies​

• Updated node dependency to v24 by @seqeralabs-renovate in #1023

Full changelog: https://github.com/seqeralabs/wave/compare/v1.33.4...v1.33.5

Compute environments​

• Added a CloudInfo cloud instance type catalog and chip-selector filters specification.
• Added a description field to the Compute Environment V2 form.

Datasets​

• Exposed dataset file size in version API responses.

Monitoring​

• Added a getCost failure metric.
• Rendered fractional Allocated CPU values on the Run Details metrics card.
• Added SCIM group audit logs.

Pipelines​

• Added a Nextflow syntax parser v2 toggle.
• Added launch-time settings to the workflow configuration tab.

Studios​

• Added the compute environment name to the Studios row card.
• Added new API endpoints to fetch compatible compute environments for a Studio and update the compute environment on a Studio.
• Enabled updating the compute environment on a Studio.
• Resolved the user on SSH Studio connection attempts.

Data Explorer​

• Cached the impersonated signer for WIF pre-signed URLs.

General​

• Introduced scroll-into-error behavior for the notification container component and unified the placement of message containers across form pages.
• Collapsed Settings breadcrumbs to two segments.
• Refined styling for the bordered variant of label values.
• Improved readability of usage chip icons on dark backgrounds.

Bug fixes​

Access control​

• Redirected on JWT expiry and stopped the onboarding polling loop.
• Auto-created organization membership during IdP group reconciliation.
• Polished the IdP group mapping UI.

Compute environments​

• Validated NVMe and Wave dependencies for Fusion v2.
• Split Forge disposal into separate resources and credentials phases.
• Ported the Forge disposal split to Azure and GCP.
• Clarified setRemoteAvailableCredits behavior and floored stale balance updates.
• Granted lineage store access to the forged AWS Batch role.
• Bumped the GCP default GPU image family to CUDA 12.9.
• Resolved the WIF credential owner context inside a read-only transaction.

Data Explorer​

• Supported GCS pre-signed URLs with WIF credentials in Data Explorer.

Pipelines​

• Disabled the lineage launch toggle when lineage settings are not configured or when the compute environment does not support lineage.
• Set the default value for Launch.syntaxParser to match the database schema.
• De-flaked the PipelineServiceTest agent revisions check.

Studios​

• Recovered legacy Studio event names in the usage query.
• Fell back to baseline Studio capabilities when the manifest is missing or the Studio version is not found in the manifest.

General​

• Pinned the Netty allocator to pooled to mitigate Micronaut 4.9 direct memory growth.
• Stopped ProviderUnrecoverableException from hiding JVM errors.
• Increased the polling timeout for the async audit image bundle test.