Ingest: Support S3 bucket in FilesGlob source

[sergiimk]

This ticket is to extend FilesGlob support to allow something like:

fetch:
  kind: FilesGlob
  path: s3://my.bucket/data/db-table-dump-*.csv
  eventTime:
    kind: FromPath
    pattern: 'db-table-dump-(\d+-\d+-\d+)\.csv'
    timestampFormat: '%Y-%m-%d'

It should also allow expressing paths to S3-like storages like Minio e.g. like:

s3+http://localhost:5050/my.bucket/data/db-table-dump-*.csv

Notes:

• Wildcards will only be supported in object path components (not hosts or bucket names)
• We can support initially patterns like (/dir/file-*.csv, /dir/**/*.json, /dir/subdir-*/*.parquet)
• This type of ingest is aimed at initial convenience only, and should be reserved for low file counts, as the operation of listing an S3 bucket may get slow/expensive at scale
• Similarly to existing FileGlob logic, the state of the source will be represented by the name of the last file ingested, with ingest proceeding in lexicographic sort order