cloud: add Table Count by TTL Schedule Delay and TTL Insert/Delete Rows by Day (#22987)

Author: huoyao1125

tidb-cloud/premium/built-in-monitoring-premium.md

@@ -43,6 +43,8 @@ The following sections illustrate the metrics on the **Metrics** page for {{{ .p
 | Transaction Duration | avg-{transaction model}, 99-{transaction model} | The average or the 99th percentile duration of transactions. |
 | Connection Count | All, active connection | All: the number of connections. <br/> Active connections: the number of active connections. |
 | Disconnection Count | {result} | The number of clients disconnected. |
+| Table Count by TTL Schedule Delay | `{name}` | Shows the number of TTL-enabled tables grouped by schedule delay. The `name` label is a delay bucket. Possible bucket values: `01 hour`, `02 hour`, `06 hour`, `12 hour`, `24 hour`, `72 hour`, `one week`, and `others`. |
+| TTL Insert/Delete Rows by Day | `insert current day`, `delete current day`, `insert last day`, `delete last day`, `insert 2 days ago`, `delete 2 days ago` | Shows the number of rows inserted into TTL-managed tables and the number of rows deleted by TTL jobs for each calendar day. A sustained gap where inserts consistently exceed deletes indicates that TTL processing is falling behind data ingestion, which can lead to storage growth. Comparing values across multiple days helps distinguish normal workload fluctuations from a long-term backlog trend. |
 
 ### Database