chore: clean up err tracker wf, ignore silenced wf

[MasterPtato]

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-4979 environment in rivet-frontend

Service	Status	Web	Updated (UTC)
kitchen-sink	❌ Build Failed (View Logs)	Web	May 9, 2026 at 11:21 pm
website	😴 Sleeping (View Logs)	Web	May 9, 2026 at 3:33 am
mcp-hub	✅ Success (View Logs)	Web	May 5, 2026 at 8:02 pm
ladle	❌ Build Failed (View Logs)	Web	May 5, 2026 at 8:00 pm
frontend-cloud	❌ Build Failed (View Logs)	Web	May 5, 2026 at 8:00 pm
frontend-inspector	❌ Build Failed (View Logs)	Web	May 5, 2026 at 8:00 pm

[MasterPtato]

• feat(vbare): add conversion boilerplate generator, fix envoy protocol conversion #5000
• chore: clean up types #4999
• chore: clean up err tracker wf, ignore silenced wf #4979 👈 (View in Graphite)
• main

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

[claude]

PR #4979 Review: chore: clean up err tracker wf, ignore silenced wf

Overview

This PR makes four targeted fixes across the workflow engine and actor runtime:

1. Gasoline KV – silenced workflows are now skipped in the main workflow poll/dispatch path
2. Guard – removes a duplicate rustls ring crypto-provider installation
3. Actor2 runtime – clears reschedule_ts on successful allocation (mirrors existing sleep_ts reset)
4. Runner pool – captures the error-tracker workflow ID and sends it a Shutdown signal when the runner pool workflow exits

---

Change-by-change notes

engine/packages/gasoline/src/db/kv/mod.rs

Good fix. The lease-expiry path at line 568 already skips silenced workflows; this change brings the main workflow-fetch path into parity. The implementation is correct:

• SilenceTsKey is read in the existing parallel tokio::try_join!, so no extra serial roundtrip is added.
• The destructuring position aligns with the new future.
• Log message uses tracing::warn! consistent with the adjacent has_output warning.
• The combined if has_output || is_silenced lease-clear block avoids duplicating the cleanup.

One minor note: silence_ts_entry is fetched on every poll iteration even for non-silenced workflows. This is fine since it is parallel with the other reads, but worth keeping in mind if silenced-workflow volume ever grows high.

engine/packages/guard/src/lib.rs

Clean removal. rivet_pools::Pools::new already calls install_rustls_provider() (pools.rs:106–108), so the call in guard::start was redundant. This is safe to drop — pools are constructed before start is entered.

engine/packages/pegboard/src/workflows/actor2/runtime.rs

Correct. When allocate finds a slot, both sleep_ts and reschedule_ts should be cleared so stale retry timers don't fire after a successful allocation. Previously only sleep_ts was reset, leaving reschedule_ts pointing at an obsolete backoff timestamp.

engine/packages/pegboard/src/workflows/runner_pool.rs

Bug fix with one migration caveat.

The error tracker runs in a lupe loop that exits only on a Shutdown signal. Before this PR, the runner pool dispatched the tracker but never sent Shutdown when it exited, leaving the tracker running indefinitely. Capturing error_tracker_wf_id and sending Shutdown at the end of the workflow is the right fix.

Migration caveat: runner pool workflows that were dispatched and completed before this code is deployed will have orphaned error-tracker workflows stuck waiting for a signal that will never arrive. These will need to be cleaned up via the silence mechanism or manual intervention. This is a timing limitation, not a flaw in the fix itself.

The ctx.v(2) version marker on the new signal send is safe. Gasoline tracks each op by its cursor position in history, so reusing version 2 for the new terminal op does not conflict with the earlier ctx.v(2).workflow(error_tracker).dispatch() call.

---

Summary

All four changes are correct and consistent with surrounding patterns. No blocking concerns. The orphaned-tracker migration side effect is the only non-obvious consequence, and it is limited in scope.