Surface critical 'events' in metrics dashboard and charts

We should provide an experience in the metrics dashboards that allow users to quickly determine when critical events such as a schema change, restarts or an upgrade occurs. This will allow a quick way to correlate critical cluster or workload events to high-level metrics patterns and behaviors.

User flow: User sees an anomaly in metrics pattern at a specific time interval. User quickly sees the event in the metrics chart for correlation. User drills into the event to determine if expected or an issue (the next step in this flow depends on the event type).

Other improvements:

- Consider custom events (time snapshots)
- Consider adding to other dashboards/charts

FYI @dongniwang

Related issue: https://cockroachlabs.atlassian.net/browse/CRDB-19805

Jira issue: CRDB-24049

Epic: CRDB-32130