fix(rivetkit): add sandbox reconnection fallback with fresh provisioning

[NathanFlurry]

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-4524 environment in rivet-frontend

Service	Status	Web	Updated (UTC)
frontend-inspector	😴 Sleeping (View Logs)	Web	Mar 30, 2026 at 11:11 pm
frontend-cloud	😴 Sleeping (View Logs)	Web	Mar 30, 2026 at 11:02 pm
website	😴 Sleeping (View Logs)	Web	Mar 30, 2026 at 10:56 pm
mcp-hub	✅ Success (View Logs)	Web	Mar 29, 2026 at 7:19 pm
kitchen-sink	❌ Build Failed (View Logs)	Web	Mar 29, 2026 at 7:18 pm
ladle	❌ Build Failed (View Logs)	Web	Mar 29, 2026 at 7:18 pm

[NathanFlurry]

• fix(rivetkit): add sandbox reconnection fallback with fresh provisioning #4524 👈 (View in Graphite)
• main

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

[pkg-pr-new]

More templates

• actor-actions
• actor-actions-vercel
• ai-agent
• ai-agent-vercel
• example-ai-and-user-generated-actors-freestyle
• chat-room
• chat-room-vercel
• example-cloudflare-workers
• example-cloudflare-workers-hono
• collaborative-document
• collaborative-document-vercel
• cross-actor-actions
• cross-actor-actions-vercel
• example-cursors
• example-cursors-raw-websocket
• example-cursors-raw-websocket-vercel
• example-cursors-vercel
• custom-serverless
• example-elysia
• example-experimental-durable-streams-ai-agent
• example-experimental-durable-streams-ai-agent-vercel
• geo-distributed-database
• geo-distributed-database-vercel
• hello-world
• hello-world-vercel
• example-hono
• hono-react
• hono-react-vercel
• example-kitchen-sink
• example-kitchen-sink-vercel
• multi-region-vercel
• multiplayer-game
• example-multiplayer-game-patterns
• example-multiplayer-game-patterns-vercel
• multiplayer-game-vercel
• native-websockets
• native-websockets-vercel
• next-js
• per-tenant-database
• per-tenant-database-vercel
• example-raw-fetch-handler
• example-raw-fetch-handler-vercel
• example-raw-websocket-handler
• example-raw-websocket-handler-proxy
• example-raw-websocket-handler-proxy-vercel
• example-raw-websocket-handler-vercel
• example-react
• example-react-vercel
• sandbox
• sandbox-coding-agent
• sandbox-coding-agent-vercel
• sandbox-vercel
• scheduling
• scheduling-vercel
• example-sqlite-drizzle
• example-sqlite-raw
• state
• state-vercel
• example-stream
• example-stream-vercel
• example-trpc

@rivetkit/cloudflare-workers

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/cloudflare-workers@4524

@rivetkit/framework-base

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/framework-base@4524

@rivetkit/next-js

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/next-js@4524

@rivetkit/react

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/react@4524

rivetkit

pnpm add https://pkg.pr.new/rivet-dev/rivet/rivetkit@4524

@rivetkit/sql-loader

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/sql-loader@4524

@rivetkit/sqlite-vfs

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/sqlite-vfs@4524

@rivetkit/traces

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/traces@4524

@rivetkit/workflow-engine

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/workflow-engine@4524

@rivetkit/virtual-websocket

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/virtual-websocket@4524

@rivetkit/engine-runner

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/engine-runner@4524

@rivetkit/engine-runner-protocol

pnpm add https://pkg.pr.new/rivet-dev/rivet/@rivetkit/engine-runner-protocol@4524

commit: bd4a98e

[claude]

PR Review: Sandbox Reconnection Fallback

The fix addresses a real pain point: stale sandbox IDs after Docker daemon restarts with AutoRemove. The two-level try/catch with state restoration is a sound approach. A few things worth thinking through.

---

Potential issue: Transient failures trigger irreversible data loss

The fallback to fresh provisioning fires on any error from SandboxAgent.start(), not just "sandbox not found" errors. A transient network blip or a momentary ECONNREFUSED while the sandbox-agent is still starting up will cause the actor to abandon the existing sandbox and provision a brand-new one, permanently losing all session history on the old sandbox.

Consider gating the fallback on an error type/message check (e.g., "not found", "does not exist") rather than catching all errors. If SandboxAgent.start() surfaces typed errors, use those. If it does not, that gap is worth filing against sandbox-agent.

---

Session re-subscription after a fresh sandbox

After the fallback path succeeds, c.state.subscribedSessionIds is empty, so the for loop at lines 232-235 is a no-op. This is correct (those session IDs no longer exist on the new sandbox), but it can be surprising when debugging. A short comment above the loop would make the intent explicit.

---

error: String(error) loses the stack trace in logs

String(error) on an Error instance gives just the message without the stack. For a warning that fires silently in production the stack is often invaluable. Consider:

error: error instanceof Error ? (error.stack ?? error.message) : String(error),

or pass the raw error object if the logger handles it natively.

---

Missing test coverage

There are no tests for the new path. Given the state mutation involved (clearing and restoring sandboxId and subscribedSessionIds), a unit test that mocks SandboxAgent.start to fail on the first call and succeed on the second would be a good regression guard.

---

Overall the logic is correct and the state-restoration on double-failure is a nice touch. The main concern is the broad catch scope that could silently replace a live sandbox on transient errors.