Add an appropriate server status metric (or multiple metrics) allowing us to attribute outages the lack of SPM-4302. That is, track agg queries that are still executing after having released their storage tickets. Such queries do not participate in execution control.

This should include:

• Counting query instances that run for more than 30ms after releasing their storage ticket(s).
• Tracking the total "non-ticketed" execution time across all such cases.