1. unify scanbytes and scan rows definition in audit log，doris
metrics，profile。 scan bytes is the uncompressed data read from the file
+ from the page cache. scan rows is the raw rows read from the file.
2. add update realtime counter interface in scanner, so other scanner
type could implement this interface to get these counters realtime.
3. the uncompressed data read in page io is wrong.
4. cputimer counter is not updated correctly.

DO NOT merge this pr to 3.0 or 3.1, because there is some behavior
change for scan bytes. If user has config workload group policy or some
monitor on scan bytes metric, there will be something wrong.

…52232 (#52584)

Cherry-picked from #52232

### What problem does this PR solve?
1. unify scanbytes and scan rows definition in audit log，doris
metrics，profile。 scan bytes is the uncompressed data read from the file
+ from the page cache. scan rows is the raw rows read from the file.
2. add update realtime counter interface in scanner, so other scanner
type could implement this interface to get these counters realtime.
3. the uncompressed data read in page io is wrong.
4. cputimer counter is not updated correctly.

DO NOT merge this pr to 3.0 or 3.1, because there is some behavior
change for scan bytes. If user has config workload group policy or some
monitor on scan bytes metric, there will be something wrong.

### What problem does this PR solve?
1. unify scanbytes and scan rows definition in audit log，doris
metrics，profile。 scan bytes is the uncompressed data read from the file
+ from the page cache. scan rows is the raw rows read from the file.
2. add update realtime counter interface in scanner, so other scanner
type could implement this interface to get these counters realtime.
3. the uncompressed data read in page io is wrong.
4. cputimer counter is not updated correctly.

DO NOT merge this pr to 3.0 or 3.1, because there is some behavior
change for scan bytes. If user has config workload group policy or some
monitor on scan bytes metric, there will be something wrong.

### What problem does this PR solve?
1. unify scanbytes and scan rows definition in audit log，doris
metrics，profile。 scan bytes is the uncompressed data read from the file
+ from the page cache. scan rows is the raw rows read from the file.
2. add update realtime counter interface in scanner, so other scanner
type could implement this interface to get these counters realtime.
3. the uncompressed data read in page io is wrong.
4. cputimer counter is not updated correctly.

DO NOT merge this pr to 3.0 or 3.1, because there is some behavior
change for scan bytes. If user has config workload group policy or some
monitor on scan bytes metric, there will be something wrong.

…53729)

### What problem does this PR solve?

Problem Summary:

### Release note

The external part is implemented according to the framework defined by
the unified audit log, Doris metrics, and scanbytes and scan rows in
Profile in #52232.

However,
- Scan bytes in the external table currently represents the bytes
counted by **the top-level File Reader** called by the scan reader
layer.
- Scan rows represents the number of scanned rows of the underlying
storage. The number of scanned rows of parquet/orc does not include the
number of rows of skipped page/rowgroup. **Note: However, there is still
a problem that `jni_reader` has not yet implemented the number of rows
that only contain the storage to be read.**

…pache#53729)

Problem Summary:

The external part is implemented according to the framework defined by
the unified audit log, Doris metrics, and scanbytes and scan rows in
Profile in apache#52232.

However,
- Scan bytes in the external table currently represents the bytes
counted by **the top-level File Reader** called by the scan reader
layer.
- Scan rows represents the number of scanned rows of the underlying
storage. The number of scanned rows of parquet/orc does not include the
number of rows of skipped page/rowgroup. **Note: However, there is still
a problem that `jni_reader` has not yet implemented the number of rows
that only contain the storage to be read.**

…pache#53729)

Problem Summary:

The external part is implemented according to the framework defined by
the unified audit log, Doris metrics, and scanbytes and scan rows in
Profile in apache#52232.

However,
- Scan bytes in the external table currently represents the bytes
counted by **the top-level File Reader** called by the scan reader
layer.
- Scan rows represents the number of scanned rows of the underlying
storage. The number of scanned rows of parquet/orc does not include the
number of rows of skipped page/rowgroup. **Note: However, there is still
a problem that `jni_reader` has not yet implemented the number of rows
that only contain the storage to be read.**

…pache#53729)

### What problem does this PR solve?

Problem Summary:

### Release note

The external part is implemented according to the framework defined by
the unified audit log, Doris metrics, and scanbytes and scan rows in
Profile in apache#52232.

However,
- Scan bytes in the external table currently represents the bytes
counted by **the top-level File Reader** called by the scan reader
layer.
- Scan rows represents the number of scanned rows of the underlying
storage. The number of scanned rows of parquet/orc does not include the
number of rows of skipped page/rowgroup. **Note: However, there is still
a problem that `jni_reader` has not yet implemented the number of rows
that only contain the storage to be read.**

forget update profile counters in update_realtime_counters
bug introduced in apache#52232

Signed-off-by: freemandealer <freeman.zhang1992@gmail.com>

forget update profile counters in update_realtime_counters
bug introduced in apache#52232

Signed-off-by: freemandealer <freeman.zhang1992@gmail.com>

forget update profile counters in update_realtime_counters. bug
introduced in #52232

…55929)

forget update profile counters in update_realtime_counters. bug
introduced in apache#52232

…55929)

forget update profile counters in update_realtime_counters. bug
introduced in apache#52232