starknet_transaction_prover: global panic hook + graceful SIGTERM shutdown

[avi-starkware]

Installs a process-wide panic hook that emits one structured event=panic
tracing event with location and backtrace, and bridges SIGTERM/SIGINT into
ServerHandle::stop so container teardown becomes visible in logs and
in-flight proofs get to finish. A second signal forces exit(1) to work
around tokio-rs/tokio#7905.

Co-Authored-By: Claude Opus 4.7 (1M context) noreply@anthropic.com

[cursor]

PR Summary

Low Risk
Changes are limited to process lifecycle, logging, and signal handling; proving and JSON-RPC request handling are unchanged.

Overview
Improves production operability of the transaction prover binary: panics and shutdowns become visible in structured logs instead of ad-hoc stderr output.

A new server/panic module installs a process-wide panic hook (after tracing init) that logs event=panic with message, source location, and a forced backtrace via tracing::error, without changing default abort-on-panic behavior. extract_payload is covered by a small unit test.

main now wires SIGTERM/SIGINT into jsonrpsee’s ServerHandle::stop, logs shutdown_started / shutdown_complete, and on a second signal logs force_exit and calls exit(1) so a stuck graceful shutdown is not swallowed (workaround for tokio signal behavior).

Dependencies: tokio gains the signal feature; tracing-subscriber gains json so the existing LogFormat::Json CLI path can emit JSON logs (including panic events).

^{Reviewed by Cursor Bugbot for commit c3950e4. Bugbot is set up for automated code reviews on this repo. Configure here.}

[avi-starkware]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• starknet_transaction_prover: document observability surface in README #14203
• starknet_transaction_prover: origin-level breadcrumbs for validation errors #14172
• starknet_transaction_prover: /health returns 503 when service is saturated #14171
• starknet_transaction_prover: bump prover_panics_total on panic hook fire #14170
• starknet_transaction_prover: HTTP request count + latency + in-flight metrics #14169
• starknet_transaction_prover: proving-job duration + outcome metrics #14168
• starknet_transaction_prover: Prometheus /metrics endpoint with build_info #14167
• starknet_transaction_prover: global panic hook + graceful SIGTERM shutdown #14166 👈 (View in Graphite)
• starknet_transaction_prover: per-request structured log with request-id #14165
• starknet_transaction_prover: redact URL credentials in logs and add startup banner #14164
• starknet_transaction_prover: add /health endpoint via HealthLayer #14163
• starknet_transaction_prover: add --log-format text|json CLI flag #14162
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[reviewable-StarkWare]

This change is 

[cursor]

Cursor Bugbot has reviewed your changes 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 c31ce65. Configure here.}

[cursor]

Reimplements already-stable standard library method

Low Severity

The extract_payload function manually reimplements PanicHookInfo::payload_as_str(), which was stabilized in Rust 1.91.0. The project's rust-toolchain.toml specifies channel 1.95, so the standard library method is already available. The doc comment on extract_payload references a nightly-2025-07-14 toolchain that no longer matches the project's actual pinned toolchain, making the stated reason for the workaround stale.

 

^{Reviewed by Cursor Bugbot for commit c31ce65. Configure here.}

[avi-starkware]

Actually, the crate starknet_transaction_prover uses the referenced nightly toolchain, so the doc comment is not stale