Set count of failure restart SMF service before goes to maintenance

I have next services, which goes to maintenance from time to time

svc:/application/pkg/system-repository:default

svc:/system/auto-update:stackdb

I can't determine the problem, because there are not any error on the local repo or any servers. So, I need a some workaround till I find a solution. I have googled a some notice (not a solution, because there aren't such properties) to use "startd/critical_failure_count" and "startd/critical_failure_period" but it doesn't work. May be there are some similar properties?

I don't want to change behaviour for all the services via "svc:/system/svc/restarter:default", I need change only for specific services.