Say, by setting a hard memory limit, and restarting it if a proc spun up by supervisord crosses the threshold?
Looked through the configuration settings and didn't see anything jump out at me in terms of doing this internally to supervisord. We've got some processes which are hogging memory, get swapped to disk, and then supervisord doesn't seem to be able to reclaim them to manage them again. Until I can figure out the ballooning memory issue I'd like to be able to have supervisord be smart enough to restart those processes when they start running away. Anyone doing this? _______________________________________________ Supervisor-users mailing list [email protected] http://lists.supervisord.org/mailman/listinfo/supervisor-users
