fix/nitro worker output root rpc

[0xOsiris]

_{Stack created with GitHub Stacks CLI • Give Feedback 💬}

---

Note

Medium Risk
Misconfigured L2 consensus URLs could break witness/output-root collection; new L1 RPC timeouts may surface failures on slow providers that previously hung silently.

Overview
Witness building no longer uses the L2 execution RPC for optimism_outputAtBlock when eth_getProof fails. OnlineHostConfig gains optional l2_consensus_rpc; Nitro/SP1 workers and the prover CLI expose --l2-consensus-rpc / L2_CONSENSUS_RPC_URL, and missing consensus RPC now fails with an explicit error instead of calling the wrong endpoint.

metered_http_client takes a per-request request_timeout (default 10s via DEFAULT_RPC_REQUEST_TIMEOUT_SECONDS). Challenger, defender, and proposer add L1_RPC_TIMEOUT_SECONDS so hung L1 connections cannot block poll loops indefinitely.

Devnet SP1 worker passes None for consensus RPC because local reth serves eth_getProof directly.

^{Reviewed by Cursor Bugbot for commit cc17e5d. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit cc17e5d. Configure here.}

[cursor]

E2E skips available consensus RPC

Medium Severity

The e2e host config passes None for l2_consensus_rpc with a comment claiming there is no op-node, but this test already requires ROLLUP_RPC_URL as the op-node for output roots. That leaves the new eth_getProof fallback unwired, so any eth_getProof failure fails the run even though a consensus endpoint is available.

 

^{Reviewed by Cursor Bugbot for commit cc17e5d. Configure here.}