[Feature] Enhancement request: Make Query.DEFAULT_CAPACITY configurable via server properties

### Feature Description (功能描述)

**Problem Statement**
The current implementation in `org.apache.hugegraph.backend.query.Query` class uses a hard-coded `DEFAULT_CAPACITY` value of 800,000 records:
```java
public static final long DEFAULT_CAPACITY = 800000L; // HugeGraph-777
```

This limitation creates the following problems:

1. When executing large graph queries that exceed this limit, users encounter the following error:
```
org.apache.hugegraph.exception.LimitExceedException: Too many records(must <= 800000) for the query: `Query * from EDGE_LABEL_INDEX where id prefix with 0x45493a2d323a4e00`
```
2. This fixed limit doesn't account for varying server capabilities - servers with substantial memory and processing power are still constrained by this arbitrary limit.
3. It forces users to implement complex client-side pagination logic for legitimate use cases that need to process more records.
4. The limit isn't documented prominently in user documentation, leading to confusion when users encounter this error.