seqera
Preview feature: may change in a future release.
The seqera scope controls the interactions with Seqera services, including the Seqera executor.
seqera.executor.autoLabels
When true, automatically adds workflow metadata labels to the session with the nextflow.io/ prefix (default: false). The following labels are added: projectName, userName, runName, sessionId, resume, revision, commitId, repository, manifestName, runtimeVersion. A seqera.io/runId label is also added, computed as a SipHash of the session ID and run name. The userName label reports the Seqera Platform user that submitted the run, falling back to the OS user running the Nextflow launcher when no Platform identity is available.
seqera.executor.computeEnvId
The Seqera Platform compute environment ID. When specified, the scheduler resolves the compute environment directly by this ID instead of inferring a suitable compute environment. Used as a fallback when the workflow launch does not include a compute environment reference.
seqera.executor.endpoint
The Seqera scheduler service endpoint URL (required).
seqera.executor.keyPairName
The EC2 key pair name for SSH access to instances.
seqera.executor.predictionModel
The resource prediction model used to estimate task resource requirements from historical execution metrics. Supported values: 'qr/v1', 'qr/v2' (quantile regression). When not set, no resource estimation is applied.
seqera.executor.providerConfig
Backend-specific provider configuration merged into the compute cluster's backend properties (for cluster isolation). When omitted, the backend falls back to its environment variable configuration. For example: providerConfig = [lazyPull: 'true'].
seqera.executor.provider
The compute backend provider type (e.g. 'aws', 'local'). When specified, used together with region to select the matching compute environment.
seqera.executor.region
The cloud region for task execution.
seqera.executor.shellEnabled
When true, enables on-demand interactive shell access (e.g. SSH) to the run's task containers for the VM and local backends (default: false). A running task can then be reached with sched task ssh <task-id> (or a plain ssh <task-id>@<scheduler>), and the connection survives task completion.
seqera.executor.strategy
The execution strategy within the chosen provider. Narrows compute-environment selection when a provider offers multiple strategies (e.g. AWS supports 'ecs' and 'vm'). When omitted, the provider's canonical default strategy is used (AWS → 'ecs').
seqera.executor.taskEnvironment
Custom environment variables to apply to all tasks submitted by the Seqera executor. These are merged with the Fusion environment variables, with Fusion variables taking precedence. For example: taskEnvironment = [MY_VAR: 'value'].
seqera.executor.machineRequirement.capacityMode
The ECS capacity provider mode. Can be 'managed' (default) for ECS Managed Instances, or 'asg' for an Auto Scaling Group-backed capacity provider.
seqera.executor.machineRequirement.diskAllocation
The disk allocation strategy. Can be 'task' (default) for per-task EBS volumes, or 'node' for per-node instance storage. When using 'node' allocation, EBS-specific options (diskType, diskIops, diskThroughputMiBps, diskEncrypted) are not applicable.
seqera.executor.machineRequirement.diskEncrypted
Enable KMS encryption for the EBS volume (default: false). Only applicable when diskAllocation is 'task'.
seqera.executor.machineRequirement.diskIops
The IOPS for io1/io2/gp3 volumes. Required for io1/io2 volume types. Only applicable when diskAllocation is 'task'.
seqera.executor.machineRequirement.diskMountPath
The container path where the disk is mounted (default: '/tmp'). Applicable to all disk allocation strategies.
seqera.executor.machineRequirement.diskSize
The disk size for session-level storage (e.g. 100.GB).
seqera.executor.machineRequirement.diskThroughputMiBps
The throughput in MiB/s for gp3 volumes (125-1000). Default: 325 (Fusion recommended). Only applicable when diskAllocation is 'task'.
seqera.executor.machineRequirement.diskType
The EBS volume type for task scratch disk. Supported types: 'ebs/gp3' (default), 'ebs/gp2', 'ebs/io1', 'ebs/io2', 'ebs/st1', 'ebs/sc1'. Only applicable when diskAllocation is 'task'.
seqera.executor.machineRequirement.machineTypes
List of acceptable EC2 instance families. For example, ['m5', 'c5', 'r5'].
seqera.executor.machineRequirement.maxSpotAttempts
The maximum number of spot retry attempts before falling back to on-demand. Only used when provisioning is 'spot' or 'spotFirst'.
seqera.executor.machineRequirement.provisioning
The instance provisioning mode. Can be 'spot', 'ondemand', or 'spotFirst'.
seqera.executor.schedulingRequirement.maxCpusPerUser
The maximum number of concurrent vCPUs a single user may consume across all of their runs on this compute environment (default: no limit). A task requesting more cpus than this cap fails instead of being queued.
seqera.executor.retryPolicy.delay
The initial delay when a failing HTTP request is retried (default: '450ms').
seqera.executor.retryPolicy.maxDelay
The maximum delay when a failing HTTP request is retried (default: '90s').
seqera.executor.retryPolicy.maxAttempts
The maximum number of retry attempts (default: 10).
seqera.executor.retryPolicy.jitter
The jitter factor for randomizing retry delays (default: 0.25).
seqera.executor.retryPolicy.multiplier
The multiplier for exponential backoff (default: 2.0).