feat(aws_s3 sink): Add Apache Parquet encoder support

[rorylshanks]

Summary

This PR adds Apache Parquet encoding support to the AWS S3 sink, enabling Vector to write columnar Parquet files optimized for analytics workloads.

Parquet is a columnar storage format that provides efficient compression and encoding, making it ideal for long-term storage and query performance with tools like AWS Athena, Apache Spark, and Presto. This implementation allows users to write properly formatted Parquet files with configurable schemas, compression, and row group sizing.

Key features:

• Complete Parquet encoder implementation with Apache Arrow integration
• YAML schema configuration support (field names → data types)
• Support for all common data types (strings, integers, floats, timestamps, booleans, etc.)
• Configurable compression algorithms (snappy, gzip, zstd, lz4, brotli)
• Row group size control for query parallelization
• Nullable field support
• Comprehensive test suite (9 unit tests)
• Full documentation for schema configuration and Parquet options

Vector configuration

sources:
  events:
    type: kafka
    bootstrap_servers: "kafka:9092"
    topics:
      - events

transforms:
  prepare:
    inputs:
      - events
    type: remap
    source: |
      parsed = parse_json(.message) ?? {}
      .uuid = parsed.uuid
      .properties = parsed.properties
      

sinks:
  s3_events:
    type: aws_s3
    inputs:
      - prepare
    bucket: my-bucket
    region: us-east-1
    compression: none  # Parquet handles compression internally

    batch:
      max_events: 50000
      timeout_secs: 60

    encoding:
      codec: parquet
      parquet:
        compression: zstd
        allow_nullable_fields: true
        schema:
          timestamp: timestamp_microsecond
          uuid: utf8
          properties: utf8

How did you test this PR?

I tested it against production Kafka data, and it produced correctly formatted Parquet files in S3.

Change Type

• Bug fix
• New feature (Parquet encoder for AWS S3 sink)
• Non-functional (chore, refactoring, docs)
• Performance

Is this a breaking change?

• Yes
• No

Does this PR include user facing changes?

• Yes. Please add a changelog fragment based on our guidelines.
• No. A maintainer will apply the no-changelog label to this PR.

References

• Closes: Support parquet columnar format in the aws_s3 sink #1374
• Closes: enhancement(aws_s3 sink): Add parquet codec #17395

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[rorylshanks]

I have read the CLA Document and I hereby sign the CLA

[thomasqueirozb]

Hey @rorylshanks, thanks for your contribution! It looks like there are failing checks (run make check-clippy for example). This is also failing to compile after merging master because you removed BatchSerializerConfig::build which is used by the clickhouse sink. I'll circle back to this PR and give it a review once I see commits pushed to this branch