Add RockDB Subcommand for printing usage per column family

[siladu]

Pretty much a lift and shift of @ahamlat's https://github.com/ahamlat/RocksdDB-Column-Families-Size

➜  besu git:(rocksdb-usage-subcommand) $BESU storage x-rocksdb usage --help
Usage: besu storage x-rocksdb usage [-hV]
Prints disk usage
  -h, --help      Show this help message and exit.
  -V, --version   Print version information and exit.
  
➜  besu git:(rocksdb-usage-subcommand) $BESU storage x-rocksdb --help
Usage: besu storage x-rocksdb [-hV] [COMMAND]
Print RocksDB information
  -h, --help      Show this help message and exit.
  -V, --version   Print version information and exit.
Commands:
  usage  Prints disk usage

$BESU --data-path=/tmp/besu storage x-rocksdb usage
****** Column family 'BLOCKCHAIN' size: 7 KiB ******
Number of keys : 70
Number of live snapshots : 0
Total size of SST Files : 4 KiB
Size of live SST Filess : 4 KiB
Column family size (with getColumnFamilyMetaData) : 4 KiB

****** Column family 'ACCOUNT_INFO_STATE' size: 1 KiB ******
Number of keys : 2
Number of live snapshots : 0
Total size of SST Files : 1 KiB
Size of live SST Filess : 1 KiB
Column family size (with getColumnFamilyMetaData) : 1 KiB

****** Column family 'TRIE_BRANCH_STORAGE' size: 1 KiB ******
Number of keys : 7
Number of live snapshots : 0
Total size of SST Files : 2 KiB
Size of live SST Filess : 2 KiB
Column family size (with getColumnFamilyMetaData) : 2 KiB

****** Column family 'TRIE_LOG_STORAGE' size: 5 KiB ******
Number of keys : 103
Number of live snapshots : 0
Total size of SST Files : 7 KiB
Size of live SST Filess : 7 KiB
Column family size (with getColumnFamilyMetaData) : 7 KiB

****** Column family 'VARIABLES' size: 1 KiB ******
Number of keys : 6
Number of live snapshots : 0
Total size of SST Files : 2 KiB
Size of live SST Filess : 2 KiB
Column family size (with getColumnFamilyMetaData) : 2 KiB

[github-actions]

• I thought about documentation and added the doc-change-required label to this PR if updates are required.
• I thought about the changelog and included a changelog update if required.
• If my PR includes database changes (e.g. KeyValueSegmentIdentifier) I have thought about compatibility and performed forwards and backwards compatibility tests

[gfukushima]

I've changed it to follow the MD format suggested by @jframe. Also have removed the experimental prefix.

[fab-10]

Like it, there are still a couple of not resolved comment

[garyschulte]

using https://github.com/vdmeer/asciitable this could be replaced by:

AsciiTable at = new AsciiTable();
at.addRow("Column Family", "Keys", "Column Size", "SST FIles Size")
at.addRule()

per cf: at.addRow(cfName, keys, columnSize, sstFilesSize);

and finally out.print(at.render());

non-blocking feedback. Using this lib might make the output a bit more flexible for different sized terminals and such

[gfukushima]

I spiked on this and didn't like the result tbh. It is sort of terminal-friendly but it did break the first column in my embedded terminal on IntelliJ. There's some customization we could use like defining the width of columns etc, but I think this might be an overkill. I'd rather have this MD formatted at the cost of having 3 output.format().