Add per-scanner dimension for scanner metrics

### Search before asking

- [x] I searched in the [issues](https://github.com/apache/fluss/issues) and found nothing similar.

### Motivation

ScannerMetricGroup scopes its metrics only by database and table, there is no per-scanner dimension. When two or more LogScanners run against the same database and table in the same client process, they all map to the same metric identity which could lead to wrong metric value to be reported.

### Solution

Add a new dimension so concurrent scanners on the same table become distinct metric series, this will allow operators to alert with scenarios, e.g. something like `max by (database, table)`, where any stuck scanner will dominate and report that value for alerting for instance.

### Anything else?

Happy to contribute if team thinks reasonable.

### Willingness to contribute

- [x] I'm willing to submit a PR!

Start by locating ScannerMetricGroup and the LogScanner entry points to understand how database and table currently form metric identity. Add the scanner dimension so concurrent scanners on the same table produce distinct metric series, then verify that the reported values and aggregation behavior distinguish each scanner.

Written by the indexing model from the issue text.