DO NOT MERGE: rivetkit test

[abcxff]

Description

Please include a summary of the changes and the related issue. Please also include relevant motivation and context.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Please describe the tests that you ran to verify your changes.

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[railway-app]

🚅 Deployed to the rivet-pr-4991 environment in rivet-frontend

Service	Status	Web	Updated (UTC)
frontend-inspector	❌ Build Failed (View Logs)	Web	May 7, 2026 at 5:12 pm
frontend-cloud	❌ Build Failed (View Logs)	Web	May 7, 2026 at 5:10 pm
kitchen-sink	❌ Build Failed (View Logs)	Web	May 7, 2026 at 5:09 pm
website	😴 Sleeping (View Logs)	Web	May 7, 2026 at 7:34 am
ladle	❌ Build Failed (View Logs)	Web	May 7, 2026 at 1:29 am
mcp-hub	✅ Success (View Logs)	Web	May 7, 2026 at 1:27 am

[abcxff]

• DO NOT MERGE: rivetkit test #4991 👈 (View in Graphite)
• fix(rivetkit): remove async drizzle db proxy #4984 : 2 other dependent PRs (#4985 , #4986 )
• main

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

[claude]

PR Review: DO NOT MERGE — rivetkit test (diagnostic tracing)

Note: This is a DRAFT diagnostic PR explicitly marked "DO NOT MERGE". The review below focuses on what would need to change before these debugging aids could ship in any form.

---

Overview

This PR adds extensive tracing instrumentation across the gateway/envoy/core request path to diagnose a bug in how in-flight requests are handled. It also includes a handful of standalone improvements:

• Manual Debug impls for ToActor, ToEnvoyMessage, and DispatchCommand
• Debug derive added to SqliteRequest
• keep_awake promise rejection downgraded from warn! to debug!
• Better public vs. internal error differentiation in lib.rs

---

Issues that must be fixed before merging

1. Debug traces logged at warn! level (critical)

All "DEBUG-TRACE:" messages in pegboard-gateway2/src/lib.rs and shared_state.rs use tracing::warn!. This poisons the WARN level with noise that has nothing to do with actionable warnings. These must be tracing::trace! or tracing::debug! — or removed entirely once the bug is identified.

// Bad — all of these should be trace!/debug!, not warn!
tracing::warn!(..., "DEBUG-TRACE: handle_request_inner entered");
tracing::warn!(..., "DEBUG-TRACE: subscribed to actor2::Stopped");
// ... (dozens more)

2. Existing debug! calls upgraded to warn! for no reason

Several legitimate tracing::debug! calls in shared_state.rs were changed to tracing::warn!, for example GC removal and message dispatch logs. These should be reverted to debug!.

3. other => catch-all arm still violates the enum exhaustiveness rule

The change from _ => to other => in the ToRivetTunnelMessageKind match in lib.rs is still a catch-all — it just names the value. Per CLAUDE.md:

Never use a _ => fall-through arm when matching on a Rust enum. Enumerate every variant explicitly.

All variants of protocol::ToRivetTunnelMessageKind should be listed explicitly so adding a new variant is a compile error instead of a silent behavior change.

---

Standalone improvements (worth keeping)

Debug impls for internal types (actor.rs, envoy.rs, task.rs): Well-written; appropriately uses finish_non_exhaustive() to omit large/sensitive fields. These can ship independently.

SqliteRequest Debug derive (sqlite.rs): Straightforward and useful.

trace! logs in actor_inner and envoy_loop: Using the correct trace! level for high-frequency per-message logs. These are appropriate as-is.

keep_awake promise rejection: warn! → debug! (actor_context.rs): Correct downgrade — a rejected keep_awake promise is not an actionable system warning.

Public vs. internal error split (lib.rs): Good intent. Public/user-facing errors at debug! avoids filling error logs with expected client errors. Two nits:

• The public-error branch drops error_chain, which was present in the original. Consider keeping it at debug! level even for public errors — it helps trace why a public error was generated.
• has_metadata = error.metadata().is_some() was silently removed from both branches. Mildly useful for log filtering; worth restoring in the internal branch.

PendingHibernation | Hibernating arm split (shared_state.rs): Splitting the combined match arm into two separate arms is logically equivalent and makes per-state tracing possible. No behavioral change.

---

Summary

Area	Verdict
"DEBUG-TRACE:" at warn! level	Must fix — revert to trace!/debug! or remove after diagnosis
Existing debug! bumped to warn!	Must fix — revert
other => catch-all arm	Should fix — enumerate all enum variants explicitly
Debug impls for internal types	Keep — good standalone addition
trace! per-message logs in envoy/actor loops	Keep
keep_awake warn → debug	Keep
Public/internal error split	Keep (minor nit on error_chain and has_metadata removal)