Similar to what we did for the spot-ce used by the job-queue we are going to create a Compute Environment which uses a couple of large instances to run the Nextflow process within a AWS Batch job.
Similar to the spot environment, we choose a name (), select the roles ( & ). As we do not need access to a local
aws-cli we can keep the standard ECS image.
In the second section we select On-Demand as the provisioning model. As this compute environment is only running supervising Nextflow processes, we do not ne and pick some
large instances as this compute environment ().
Furthermore, we are going to to choose the
BEST FIT PROGRESSIVE () as it will react to failures to acquire capacity.
We leave most of the defaults, set the minimal and desired vCPUs to 1 .
We keep the defaults, add a tag so that we can identify the compute-environment in the third section and click create .