chore: disabled get_batch consensus RPC

[pompon0]

get_batch is not fully implemented yet (neither in era-consensus nor zksync-era) but it already is consuming resources (db queries and bandwidth). We should finish implementation and loadtest it before enabling it again.

[aakoshh]

This looks reasonable to me, if the traffic it generates is unrealistic. I thought it might be indicative of the volume of data it would download, maybe worth having a look at the metrics even if we can't actually evaluate the proof and all we do is discard the batches.

Just to give some extra context, IIUC the batches are pulled because run_batch_fetcher gets them, always trying to download one more than what it already has in its queued cache. However I was never able to ascertain how things make their way into the queued at the moment, which could be because this hasn't been developed yet.

Did this actually generate traffic already?

[pompon0]

yes, you can see that get_batch RPC takes up to 3s https://grafana.matterlabs.dev/d/eba66cd9-5335-4521-9777-ad34a379f35c/era-consensus?orgId=1&var-datasource=mIPIusa4k . I haven't analyzed the exact impact on the db resources, but I'd like to understand it before we enable it on testnet or mainnet.