crosscluster/physical: expose reader tenant reader time in SHOW TENANT WITH REPLICATION STATUS

Currently, a PCR user has no visibility into the historical time that the _reader tenant_ uses to query the app tenant. In the happy case, it's approximately equal to the replicated time, but it can fall behind if the reader tenant job poller job stalls for some reason.

As a very first step, we should update the standby poller job's resolved time in the `system.job_progress` table so a user logged into the reader tenant's db console can observe it.

However, we should strive to simplify the PCR UX such that all PCR observability can be viewed from the system tenant, via SHOW TENANT WITH REPLICATION STATUS and system tenant time series metrics.

Jira issue: CRDB-51634

Epic CRDB-52290

Start by tracing the standby poller job and its resolved time in system.job_progress, then inspect the SHOW TENANT WITH REPLICATION STATUS entry point. Confirm how the reader tenant's time is represented and how replication status is exposed before deciding what observability should be added. Done means the reader tenant's historical query time is visible as specified, with coverage for the reported status.

Written by the indexing model from the issue text.