When running a large workflow (with many jobs >1000), and running at the same time some jobs on the cluster using arv-mount with a large collection, arvados-ws process grows memory usage significantly over 8GB. This has triggered several times the oom-killer
Updated by 
Updated by 
Updated by