fix: increase timeouts on transaction querying on L1

[montekki]

Digging into recent panics in prod it looks like they are just the case of insufficient timeouts on L1 transaction fetching.

[RomanBrodetski]

@montekki can you please quickly summmarize why this is necessary?

[montekki]

@RomanBrodetski so the race in question here is on https://matter-labs.app.eu.opsgenie.com/alert/detail/2d02ea8d-0a36-4c8e-a273-56c5fff79742-1695969365058/details, racing on transaction https://etherscan.io/tx/0x0f5171baf1c5537e09606cc527cdca8b3efd2684b731791a442a342332606fd4. The timestamps on both suggest that transaction went in at 6:35:35 UTC, our panic happened in 6:36 (+some seconds).

Furthermore this transaction went in ok, was not a part of any block reorg, currently it belongs to a finalized block. As such I see no other reason for the panic above to happen other than insufficient timeouts on the transaction querying. Before this patch panic happened if the TX couldn't be queried from RPC in 15 seconds (intuitively block time), however this limit looks insufficient looks like.

It would probably make sense to make this configurable however at this moment i don't think it is worth adding a new config var.