signpost mark-successful-boot needs to run at boot

[iliana]

• Figure out when in the boot process this should run
• Write a systemd unit

[iliana]

@jahkeup Do you have any thoughts on how this might integrate with k8s, or should we just (for now) implement a naïve version that runs after systemd believes everything is in a good state?

[jahkeup]

Let's go with the naive version, it's unlikely that dogswatch will do anything other than interpret the state of the node based on the host's reporting - if there's going to be handling for cluster-wide impact we'll want to teach the operator some new tricks (like watching Pod stability on an updated node after an update before proceeding to update other nodes).