Consider a nextflow workflow that uses the Google Life Sciences API and uses preemptible machines. The config might look like this:
google {
project = "cool-name"
region = "cool-region"
lifeSciences {
bootDiskSize = "200 GB"
preemptible = True
}
}
Let's say you have only a single process and this process has the directive maxRetries = 5. After five retries (=after the 6th time), the process will be considered failed.
Is it somehow possible to specify in nextflow that after a certain number of retries were unsuccessful, that nextflow should request a non-preemptible machine instead and continue retrying a couple more times?