Idea from @chasemp in a labs-admin email thread:
I really don't want to call these target metrics or anything but "total instances in Cloud VPS", "Total tools in Tools", "% on Grid", "% on k8s", "new ldap user count", "new Tools user count", "nova-fullstack failures (? or uptime?)"....if this was run on a cron and sent via email to the labs-admin list before the meeting that would be ideal I guess.
We have a similar weekly or months stats generator for Phab (user count, issue count, etc) and over time it creates a view of normal. If we suddenly have 100 new ldap users where we usually have 10, or if someone notices new Tools hasn't gone up in a few weeks, etc. These are hard to quantify baselines where we will be happy to we put them in front of our faces over time I believe.