Add metrics and reduce log verbosity in iceberg-source

[lawofcycles]

Description

Adds operational metrics to the iceberg-source plugin and reduces log verbosity for high-frequency messages, following the patterns established by the DynamoDB and RDS source plugins. This is a follow-up to #6554.

Metrics added:

Metric	Type	Description
changeEventsProcessed	Counter	CDC events written to buffer
changeEventsProcessingErrors	Counter	CDC partition processing failures
exportRecordsProcessed	Counter	Initial load records written to buffer
exportRecordsProcessingErrors	Counter	Initial load processing failures
snapshotsProcessed	Counter	Iceberg snapshots fully processed
bytesProcessed	DistributionSummary	Data file bytes read
carryoverRowsRemoved	DistributionSummary	Carryover rows removed per task

Log level changes:

Per-partition and per-file log messages (e.g. "Processing partition...", "Reading file...", "Carryover removal...") are changed from INFO to DEBUG. INFO is now reserved for lifecycle events (start, stop) and snapshot-level progress in LeaderScheduler.

Refactoring:
Extracted IcebergDataFileReader from ChangelogWorker to decouple Iceberg file I/O (Parquet/Avro/ORC reading) from processing logic. This enables unit testing of metrics recording by allowing test substitution of the file reader.

Issues Resolved

Related to #6552

Check List

• New functionality includes testing.
• New functionality has a documentation issue. Please link to it in this PR.
• New functionality has javadoc added
• Commits are signed with a real name per the DCO

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[lawofcycles]

I have opened #6682 which adds source-layer shuffle to the iceberg-source plugin. It modifies ChangelogWorker, LeaderScheduler, and TaskGrouper significantly. Once #6682 is merged, this PR will need a rebase to resolve conflicts, particularly around the IcebergDataFileReader extraction and the SHUFFLE_WRITE/READ processing added to ChangelogWorker.