Nimbus VC doesn't prune the slashing db on holesky

[yokem55]

Describe the bug
The nimbus_validator_client doesn't appear to prune old attestations out of the signed_attestations table in the slashing-protection.sqlilte3 db on holesky node + vc with 5K validators. The database then has unbounded growth, reducing performance of attestations over time.

To Reproduce
Steps to reproduce the behavior:

1. Platform details (OS, architecture): linux amd64
2. Branch/commit used: Nimbus validator client v23.10.0-8b07f4-stateofus
3. Commands being executed: /usr/local/bin/nimbus_validator_client --non-interactive --beacon-node=http://127.0.0.1:5052 --block-monitor-type=event --data-dir=/srv/holesky_data/vc-1 --validators-dir=/srv/holesky_data/keys/ --secrets-dir=/srv/holesky_data/secrets/ --doppelganger-detection=false --suggested-fee-recipient=0xAAAAAAA0 --metrics --metrics-address=0.0.0.0 --metrics-port=8121 --graffiti=Stuff
4. Relevant log lines: N/A. But for a given validator index in the slashing db (1), there are currently 267 epochs in the db for the validator since I last manually pruned the slashing db with dark magic that shouldn't be discussed in public. It repeats even when I 'safely' shutdown the vc, delete the slashing db all together, wait 2 epochs, and start up again.

sqlite> select count(validator_id) from signed_attestations where validator_id = 1;
267

Screenshots
N/a

Additional context
This is a very large holesky node with 5k keys. It's possible I guess that the large number is contributing to this? I can break a few of the validators off to run in a different vc instance if desired for testing.

[tersec]

#5551