Failed sub job is not catched by the main job

My illumina_qc workflow 2823 has a blue 🔩 (écrou). NG6 squeue says there is only the monitoring job alive. In the hidden directory /work/ng6/jflow/work/illumina_qc/wf002823/.working/e125805cfa, there is a slurm.status file:

/work/ng6/jflow/work/illumina_qc/wf002823/.working/e125805cfa $ ll
total 770K
-rw-r--r-- 1 ng6 NG6 191K Aug 29 22:30 Makeflow
-rw-r--r-- 1 ng6 NG6 366K Aug 29 23:48 Makeflow.makeflowlog
-rw-r--r-- 1 ng6 NG6   17 Aug 29 22:30 slurm.status.7509032
-rwxr-xr-x 1 ng6 NG6  299 Aug 29 22:30 slurm.wrapper
drwxr-xr-x 3 ng6 NG6 4.0K Aug 29 22:30 _Stash

but the associated job is cancelled for out of memory

seff 7509032
Job ID: 7509032
Cluster: genobull
User/Group: ng6/NG6
State: CANCELLED (exit code 0)
Cores: 1
CPU Utilized: 04:26:42
CPU Efficiency: 90.79% of 04:53:46 core-walltime
Job Wall-clock time: 04:53:46
Memory Utilized: 2.93 GB
Memory Efficiency: 100.00% of 2.93 GB

-> Erros was not catched, this is a problem!