Don't use multiple worker threads for integration

It appears this triggers some weird bug when used in combination with
bwa-mem2, which causes the alignment rule to crash randomly.
6 jobs for speedup in 98 minutes and 7 seconds (queued for 1 second)
Status Name Job ID Coverage
  Sanity
passed test_sanities #7124
docker

00:00:22

 
  Dry Run
passed test_dry_run #7125
docker

00:00:21

 
  Integration
failed test_integration #7142
slurm

01:37:23

failed test_integration #7129
slurm

00:00:04

failed test_integration #7126
slurm

01:05:00

failed test_integration #7132
slurm

01:33:29

 
Name Stage Failure
failed
test_integration Integration
No job log