Improve track failure reason names

[corverroos]

🎯 Problem to be solved

The tracker failure reasons that we instrument and describe here use names linked to the components or steps where the duty failed. This results in unintuitive logs and metrics which is hard for users to understand and require lookup in the reference.

We could improve operator UX by renaming these reasons to be more descriptive of the cause where applicable.

🛠️ Proposed solution

Few things we can do:

• Prefix "bug" to reasons that are unexpected, also put this in a separate section in the docs.
• If there is only one cause, use that as the reason, not the step.

Some suggestions:

• bcast -> broadcast_bn_error
• chain_inclusion -> not_included_onchain
• consensus -> no_consensus
• fetcher_aggregator_failed_prepare -> failed_aggregator_selection
• fetcher_aggregator_few_prepares -> insufficient_aggregator_selections
• fetcher_aggregator_no_external_prepares -> no_aggregator_selections
• fetcher_aggregator_no_att_data -> missing_aggregator_attestation
• fetcher_bn -> fetch_bn_error
• fetcher_proposer_failed_randao -> failed_proposer_randao
• par_sig_ex_receive -> no_peer_signatures
• par_sig_db_insufficient -> insufficient_peer_signatures
• validator_api -> no_local_vc_signature

Bugs:

• agg_sig_db -> bug_aggregation_error
• duty_db -> bug_duty_db_error
• fetcher_error -> bug_fetch_error

[boulder225]

since it introduces a breaking change, we prefer to revamp it in V2.

[gsora]

Closed by #3033