NDS Labs currently allow people to run services that remain online indefinitely. We should probably discuss the longevity of inactive services and how we plan to handle that.

Options might include:

• Timeout user services by default, but allow user to specify "indefinite" case
• Add an optional timeout field to the spec, allowing us to tune the default timeout for each service

This ticket is complete when we have discussed the necessity and possibilities for application timeout and have filed any additional work for work that we deem necessary.

Aside: Should we attempt to detect inactivity?
We would need to decide what it means for a user to be "inactive":

• Account hasn't logged in?
• Namespace has not changed?
• Endpoint / ingress hasn't been hit?