Implement datasource metrics

[kortemik]

Description
Implement datasource metrics

Use case or motivation behind the feature request
Currently users do not know the progress of a query which is frustrating. This needs to be fixed. Now that the datasource uses Spark 3 APIs it is possible to provide metric information about the datasource progress.

Please create at least following metrics aggregated into JSON data format:
Driver:

• Current archive offset
• Kafka offset

Task

• Amount of records processed
• Amount of bytes processed
• Bytes per second
• Records per second

Please consider implementing a precreated (hourly/automatic) buckets within the driver for earliest-latest span and binning the processed data in the tasks into these created buckets.

Please define JSON schema once initial development is done.

Related issues
teragrep/ajs_01#70 depends on this

Additional context
See example at #74 and close when implemented.

[kortemik]

this feature replaces "metricsLogger" in DPLParserCatalystContext on pth_10.